tag:blogger.com,1999:blog-67488774436992900502024-03-17T23:13:48.783-07:00eMpTy PagesComments about translation technology, new collaboration models, and inspirationKirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.comBlogger25319tag:blogger.com,1999:blog-6748877443699290050.post-87559950431898005092023-12-08T13:33:00.000-08:002023-12-08T13:36:18.505-08:00An Overview of ModernMT V7<p> Serious MT technology development requires ongoing efforts and
research to continually improve the performance of systems and to
address important emerging requirements as the use of MT expands.
Researchers have been working on MT for over 70 years and success
requires a sustained and continuing effort.</p><h3 style="text-align: left;"><b><span style="font-size: medium;">These efforts
approach the goal of producing as close as possible to human-quality MT
output in multiple ways, and these improvement strategies can be
summarized in the following ways:</span></b></h3><ol><li>Acquire <strong>better and higher volumes of relevant training data.</strong>
Any AI initiative is highly dependent on the quality and volume of the
training data that is used to teach the machine to properly perform the
task.</li><li>Evaluate <strong>new algorithms that may be more effective</strong>
in extracting improved performance from available training data. We
have seen the data-driven MT technology evolve from Statistical MT (SMT)
to various forms of Neural MT (NMT) using different forms of deep
learning. The Transformer algorithm which also powers LLMs like GPT-4 is
the state-of-the-art in NMT today.</li><li>Use <strong>more powerful computing resources</strong>
to dig deeper into the data to extract more learning. As the demand for
translation grows with the massive increases in content and
ever-expanding volumes of user-created content (UGC) it becomes
increasingly important for MT to handle massive scale. Today there are
global enterprises that are translating billions of words a month into a
growing portfolio of languages and thus scalability and scale are now
key requirements for enterprise MT solutions. Some researchers use more
computing during the training phase of the MT model development process
as there can be quality advantages gained at inference from doing this
extra-intensive training. </li><li>Build <strong>more responsive and integrated human-machine collaboration processes </strong>to
ensure that expert human feedback is rapidly incorporated into the core
data used to tune and improve these MT engines. While the benefits
gained from more and better data, improved algorithms, and more
computing resources are useful, t<strong>he integration of expert human
feedback into the MT model's continuous learning is a distinctive
advantage that allows an MT model to significantly outperform models
where only data, algorithms, and compute are used.</strong></li><li><strong>Add special features</strong>
that address the unique needs of large groups of users, or use cases
that are being deployed. As the use of MT continues to build momentum
with the enterprise many specialized requirements also emerge e.g.
enforcement of specific terminology for brand integrity, profanity
filters to avoid egregious MT errors, and improvement of
document-specific content awareness.</li></ol><p>All these different
approaches have the goal of producing improved MT output quality and it
will require progress along all of these different fronts to achieve the
best results. </p><h1 style="text-align: left;"><span style="color: #2b00fe;">The
ModernMT development team pursues ongoing improvements along all these
fronts on an ongoing basis, and ModernMT V7 is the result of several
measured improvements on many of these dimensions to provide improved
performance. </span></h1><p></p><p>As machine translation (MT) continues to
evolve and expand beyond the traditional use case areas such as
e-commerce, global collaboration, and customer care, those interested in
the expanding future of localization are now also looking to use
generative artificial intelligence (AI) and, in particular, large
language models (LLMs) such as OpenAI’s GPT</p><p><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgtyv5789L4AS0ZwEiQj3LcdZsgNJAsDLf8F-UB3luo0-1X_AUp-HaORcdcUnaUhCE0dR0-DCyhPjQUV1HiDgciEwUVcXB1l6lV9S_inks4m33aEYGOlRAF3D1suefkE-bxhoDTWwib2mbC3NVhTMYspkjaasxftHc8hi5mXIyjRQOD-0ujT8N9O3IUzL-o/s1600/Senza-titolo-2.png" style="margin-left: 1em; margin-right: 1em; text-align: center;"><img border="0" data-original-height="680" data-original-width="1600" height="272" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgtyv5789L4AS0ZwEiQj3LcdZsgNJAsDLf8F-UB3luo0-1X_AUp-HaORcdcUnaUhCE0dR0-DCyhPjQUV1HiDgciEwUVcXB1l6lV9S_inks4m33aEYGOlRAF3D1suefkE-bxhoDTWwib2mbC3NVhTMYspkjaasxftHc8hi5mXIyjRQOD-0ujT8N9O3IUzL-o/w640-h272/Senza-titolo-2.png" width="640" /></a></p><p>Unlike typical Neural MT, LLMs prioritize fluency over accuracy. But
while LLMs show promising results in improving the fluency of
translations, they can also produce confabulations (hallucinations),
i.e. output that is inaccurate or unrelated to the input data and thus
require careful monitoring and oversight to ensure accuracy.</p><p>With the <a href="https://translated.com/modernmt-7-with-trust-attention?ref=blog.modernmt.com" rel="noreferrer noopener">latest release of ModernMT (V7)</a>, Translated has introduced <strong>a novel technique to increase the accuracy of neural MT models</strong>, called “Trust Attention,” which can also be used to <strong>address reliability within generative AI models</strong>.</p><h3 style="text-align: left;"><span style="color: #2b00fe;">The
design and implementation of Trust Attention was inspired by how the
human brain prioritizes trusted sources in the learning process, linking
the origin of data to its impact on translation quality.</span></h3><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhYWZo3-MNXhn-Gyy-s-ysRMjLd1SfWBoEZPCjPIl4ZvQkr6TxOllT48fj_nqaLgXurRxs53zypSMUGfr8FpEy4cB3zm-Evw9inuGpq8Kw8flDcynb_FeKdMwsA83YojTrIsY-ViXXifRj3GdqKdrSjraatBRnibivp0CS8JkgTN6HLWdCNuvF9Og7hdykS/s1600/trust-attention-to-boost-quality-01.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="457" data-original-width="1600" height="114" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhYWZo3-MNXhn-Gyy-s-ysRMjLd1SfWBoEZPCjPIl4ZvQkr6TxOllT48fj_nqaLgXurRxs53zypSMUGfr8FpEy4cB3zm-Evw9inuGpq8Kw8flDcynb_FeKdMwsA83YojTrIsY-ViXXifRj3GdqKdrSjraatBRnibivp0CS8JkgTN6HLWdCNuvF9Og7hdykS/w400-h114/trust-attention-to-boost-quality-01.png" width="400" /></a></div><br /><div><p>ModernMT V7 preferentially uses the most trusted data (identified by
users) and thus the highest quality and most valuable training data has
the greatest influence on how a model performs. This is in stark
contrast to most MT models which have no discernment of data quality and
thus tend to perform using only statistical density as the primary
driver of model performance. </p><p>The Trust Attention capability
prioritizes its learning based on data value and importance like how humans sift through multiple sources of information to
identify the most trustworthy and reliable ones. Data extracted from
translations performed and reviewed by professional translators is
always preferred over other data, especially unverified translation
memory content acquired from web crawling, which is typically used by
most MT systems today.</p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">The
development team at ModernMT considers Trust Attention to be as
significant an innovation as Dynamic Adaptive MT engines. It is the
kind of feature that can dramatically improve MT system performance for
different use cases when properly used.</span></strong></h3><p></p><p>According to an evaluation by professional translators, done to validate the beneficial impact, <a href="https://blog.modernmt.com/modernmt-introduces-trust-attention-to-improve-mt-quality/" rel="noreferrer noopener">Trust Attention alone improves MT quality by up to 42%</a>,
and by an average of 16.5% in cases across the top 50 languages.
Interestingly, even many high-resource languages, such as Italian and
Spanish, showed significant improvements (in the 30% range) in human
evaluations.</p></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPyU-R8iWG04gQMFuRN9rXxKnX_M9x1vfoOFQFLhDm76PhLusnbC-yxbPdvx-gfzIDfl2Qt1JwAp4QxTiUQGBthyphenhyphenVomGmf1NXShOHjL1c_1Ta7R-pMHTpXn-gLyjkEbWGAbv4G1r4q_55lYh13EuZKou3Kb8o7YBrcqdGbr7mgczwphwW-q7EeElyBYdiw/s1600/trust-attention-to-boost-quality.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="557" data-original-width="1600" height="139" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjPyU-R8iWG04gQMFuRN9rXxKnX_M9x1vfoOFQFLhDm76PhLusnbC-yxbPdvx-gfzIDfl2Qt1JwAp4QxTiUQGBthyphenhyphenVomGmf1NXShOHjL1c_1Ta7R-pMHTpXn-gLyjkEbWGAbv4G1r4q_55lYh13EuZKou3Kb8o7YBrcqdGbr7mgczwphwW-q7EeElyBYdiw/w400-h139/trust-attention-to-boost-quality.png" width="400" /></a></div><br /><h3 id="modernmt-v7-new-features-up-to-60-better-mt-quality"><strong>ModernMT V7 New Features: Up to 60% Better MT Quality</strong></h3><p>ModernMT V7 is the evolution of Translated’s renowned adaptive MT system, recognized as a <a href="https://translated.com/machine-translation-leader-IDC?ref=blog.modernmt.com" rel="noreferrer noopener">leader in the Machine Translation Software Vendor Assessment</a>
for enterprises by IDC Marketscape 2022, and as “the most advanced
implementation of responsive MT for enterprise use” in CSA Research’s
2023 <a href="https://csa-research.com/ModernMT?ref=blog.modernmt.com" rel="noreferrer noopener">Vendor Briefing</a>.</p><p>In addition to Trust Attention, ModernMT V7 includes several other new features that <strong>further enhance the reliability and dependability of MT output</strong>. Here are the most impactful:</p><ul><li><a href="https://translated.com/brand-specific-terminology-in-modernmt?ref=blog.modernmt.com" rel="noreferrer noopener"><strong>Advanced Terminology Control</strong></a>: Along with its ability to learn the client’s terminology from past translations, ModernMT now provides companies with <strong>self-managed glossary control to ensure brand and context-specific terminology consistency</strong>.
This ability to enforce terminology has not been needed in the past
because the dynamic adaptive MT technology acquires terminology very
effectively even without this feature.</li><li><strong>DataClean AI</strong>: V7 relies on a new sanitization algorithm that identifies and removes poor-quality data to refine the training material and <strong>reduce the likelihood of hallucinations</strong>.
The close examination of errors over many years has provided clues on
the root causes of strange output from MT engines. This learning and
related benefits also transfer to LLM-based MT engines should they
become more viable in the future.</li><li><a href="https://translated.com/expanded-document-context?ref=blog.modernmt.com" rel="noreferrer noopener"><strong>Expanded Context</strong></a>:
ModernMT can now leverage up to 100,000 words of document content —Four
times more than GPT-4 - to preserve style and terminology preferences, <strong>providing unparalleled document-specific accuracy</strong> in MT suggestions and providing controls to solve persistent problems such as <strong>gender bias and inconsistent terminology</strong>.</li><li><strong>Profanity Filter</strong>: V7 masks words in translation suggestions that could be regarded as inappropriate in the target language, <strong>minimizing the possibility of cultural offenses</strong>.</li></ul><div><br /></div><h3 style="text-align: left;"><span style="color: #2b00fe;">The
combined effect of all the improvements and innovations described above
has a significant impact on the overall performance and capabilities of
ModernMT.</span></h3><h3 style="text-align: left;"><span style="color: #2b00fe;">The MT quality is now considered to be </span><strong><span style="color: #2b00fe;">45% to 60% better than th</span>e<span style="color: #2b00fe;"> previous version according to systematic human evaluations</span></strong><span style="color: #2b00fe;">.</span></h3><div><span style="color: #2b00fe;"><br /></span></div><p></p><p>These
improvements have greatly reduced the Time to Edit (TTE) for MT
suggestions. At the end of July, the aggregate TTE measured across tens
of thousands of samples showed a 20% reduction, reaching a record low of
1.74 seconds. This milestone indicates an acceleration towards <a href="https://translated.com/speed-to-singularity?ref=blog.modernmt.com" rel="noreferrer noopener">singularity in translation</a>, a trend further supported by preliminary TTE data collected continuously since the 1.74 seconds record was established.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgFJCumBvrYfTQHyMR2GiDD92N7HiTTe1wj_sMvOhxNE96AISgYFm7ROh-LZXDjkTrw8KuDTDJx8cqSxQnGLlCf7K_1dWQAAukkEA944SGhl8UMsDnUy_mz-ps6uPQFtybLC721YN3EoJxm7DvovniAKkJGZcitaFejMW0OSGBxHDcZ6CfYvLkHgiaf2iu1/s1999/Senza-titolo-2-copia.jpg" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="941" data-original-width="1999" height="189" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgFJCumBvrYfTQHyMR2GiDD92N7HiTTe1wj_sMvOhxNE96AISgYFm7ROh-LZXDjkTrw8KuDTDJx8cqSxQnGLlCf7K_1dWQAAukkEA944SGhl8UMsDnUy_mz-ps6uPQFtybLC721YN3EoJxm7DvovniAKkJGZcitaFejMW0OSGBxHDcZ6CfYvLkHgiaf2iu1/w400-h189/Senza-titolo-2-copia.jpg" width="400" /></a></div><h1 style="text-align: left;"><strong>The Hallmark of the Symbiosis Between Translators and MT</strong></h1><p>ModernMT V7 is <strong>available in 200 languages</strong> and <a href="https://blog.modernmt.com/modernmt-significantly-expands-language-coverage/" rel="noreferrer noopener">covers all the fastest-growing economies</a>
likely to emerge over the next 20 years. Its hallmark is the ability of
the MT model to learn from corrections in real time, enabling a <strong>powerful collaboration between the expertise of professional translators and the speed and capacity of MT</strong>. </p><p>Thanks
to this unique approach, combined with Translated’s vast community of
professional translators and leading AI-enabled localization solutions (<a href="https://translated.com/gartner-market-guide-ai-translation-services-2022?ref=blog.modernmt.com" rel="noreferrer noopener">Gartner 2022</a>), Airbnb was able to <a href="https://drive.google.com/file/d/1ob9AHgqRFJc_vkNK6JKbYgk2TTf97be7/view?usp=sharing&ref=blog.modernmt.com" rel="noreferrer noopener">ditch the translate button</a>
and simply make multilingual content pervasive and comprehensive across
the platform and become one of the top 3 global brands (<a href="https://globalbydesign.com/2023/02/25/the-best-25-global-websites-from-the-2023-web-globalization-report-card/?ref=blog.modernmt.com" rel="noreferrer noopener">Global by Design 2023</a>).</p><p>Success stories like that of Airbnb and others, along with market
research that shows the ever-growing demand for more multilingual
content, have led Translated to estimate that once MT reaches what is
commonly referred to as “parity with human translation” (<a href="https://translated.com/singularity-in-AI-impact-on-translation-industry?ref=blog.modernmt.com" rel="noreferrer noopener">singularity in translation</a>), we can expect <strong>a 100-fold increase in MT requests alongside a 10-fold growth in demand for professional translations</strong>.</p><p>We
are entering a new era in which significantly larger volumes of content
will be translated automatically. In this scenario, professional
translators play an increasingly important role, not only in guiding the
MT through the adaptive process but also in ensuring that the key
messages are appropriately conveyed. By engaging the best translators
with the best adaptive MT, companies can now take on projects that
simply weren’t feasible before.</p><h3 id="towards-llms-for-translation" style="text-align: left;"><strong>Moving Towards LLMs for Translation</strong></h3><p>Recently, Translated conducted a <strong>large-scale study to compare the performance of the most advanced MT systems with LLMs in terms of enterprise readiness</strong>.
The findings showed real potential for LLMs, particularly in terms of
more fluent translation quality, and also revealed areas where
improvements are needed. Based on this research, Translated believes
elements of <strong>both MT systems and LLMs will be critical as we move forward</strong>, and plans to provide in-depth insights into using LLMs in translation in the coming weeks and months.</p><p>Comments by John Tinsley of Translated SRL on LLM-based Translation in November 2023:</p><p><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">❗ LLMs - the new default for machine translation ❗</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">I've seen a lot of commentary along these lines over the past few months. I've also seen a lot of well-articulated commentary, not strictly opposing this line, but with added nuance and context (a challenge on the internet!)</span><span class="white-space-pre" color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline); white-space: pre;"> </span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">I wanted to offer my two cents, from being at the forefront of these developments through actually building the software, and from having many conversations with clients.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">In summary, today, LLMs are not fit for purpose as a drop-in replacement for MT for enterprises.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">More broadly, any general-purpose GPT application will find it super challenging to outperform a purpose-built enterprise solution that considers an entire workflow in a holistic way (note, the purpose-built solution could be GPT-based itself, but with a much narrower scope).</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">🧠 As a concrete example, at Translated, we've built a version of ModernMT that uses GPT-4 as a drop-in replacement for our Transformer model (while retaining the framework in ModernMT that allows us to do real-time adaptation). We've also built, and continue to test, a version of ModernMT with other open source LLMs fine-tuned for translation.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">While we find that they perform well in terms of quality on some content types and some languages, it's far from unanimous across the board. And that's just quality. Other critical enterprise factors such as speed, cost, and importantly, information security, are just not there yet. Similarly, language coverage for LLMs is a challenge as there are large discrepancies in performance, particularly for content generation.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">I appreciate there's a lot of downward pressure today to use AI across workflows, particularly in localization teams for translation and content creation. Let me hop on my soapbox to give you some information that might help with those conversations...</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">📣 If you're using MT, you're already using very advanced AI! 📣</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">You probably already know that the T in GPT stands for Transformer. But did you know that the Transformer was invented at Google in 2017...specifically for machine translation!? So what we're seeing today is a repurposing of that technology for a different application (generative AI) other than translation.</span><span class="white-space-pre" color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline); white-space: pre;"> </span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">There will come a day, possibly soon, when it's better across the board to use LLMs for translation. When that happens, it will become the standard and people will stop talking about it. Just like when Neural MT came on the scene ~6 years ago.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; font-size: 14px;">When it happens, Translated will have already deployed it in ModernMT and worked out the best way for you to adapt it to your business. We already have a lot of ideas. We already have a lot of data from the testing I mentioned earlier. And in the meantime, we still have what I believe to be the most complete enterprise translation solution available.</span><span color="rgba(0, 0, 0, 0.9)" face="-apple-system, system-ui, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", "Fira Sans", Ubuntu, Oxygen, "Oxygen Sans", Cantarell, "Droid Sans", "Apple Color Emoji", "Segoe UI Emoji", "Segoe UI Emoji", "Segoe UI Symbol", "Lucida Grande", Helvetica, Arial, sans-serif" style="background-color: white; border: var(--artdeco-reset-base-border-zero); box-sizing: inherit; font-size: 14px; line-height: inherit; margin: var(--artdeco-reset-base-margin-zero); outline: var(--artdeco-reset-base-outline-zero); padding: var(--artdeco-reset-base-padding-zero); vertical-align: var(--artdeco-reset-base-vertical-align-baseline);"><br style="box-sizing: inherit; line-height: inherit;" /></span></p><p><br /></p><p><br /></p><figure class="kg-card kg-image-card kg-width-full"><br /></figure>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0tag:blogger.com,1999:blog-6748877443699290050.post-56637582727753179232023-12-07T17:38:00.000-08:002023-12-07T17:39:51.655-08:00Prioritization of Trustworthy Data in NMT Model Development<p> </p><h3 style="text-align: left;">ModernMT: A History of Innovation and Evolution</h3><p>Neural
machine translation (NMT) has had impressive evolutionary progress over
the last five years, showing continually improving performance in
accuracy. This progress is specially marked and clear with the
dynamically adaptive NMT models like ModernMT, where small amounts of
ongoing corrective expert feedback results in continuously improving MT
output quality.</p><p>The historical track record with ModernMT has been
so impressive that it did not seem unreasonable to point out that
ModernMT's performance across billions of samples and many languages
was <a href="https://blog.modernmt.com/the-march-towards-singularity/">approaching singularity in production-use</a>
scenarios. This is a point at which human editors are unable to tell
whether the sample is coming from a human or machine since they are so
close in quality and style.</p><p>NMT technology continues to evolve and
improve with recent updates that provide much richer and more granular
document-level contextual awareness. Document-level adaptation in
machine translation has been a core design intention with ModernMT from
the outset. This originally involved referencing similar sentences in
translation memories and using these to influence new translation
requests. </p><p>Despite the success and pioneering nature of this
approach, early implementations faced challenges: translators struggled
with issues such as gender bias and inconsistent terminology due to the
distance between the segment they were working on and its related
context.</p><p>By taking into account all edits within an individual
document, even those in completely different or distant segments, the MT
model is now able to provide document-specific translation suggestions.
This development significantly reduces the need for repeated
corrections of elements such as pronouns. This has greatly eased the
amount of corrective work needed to address gender bias errors and
modify incorrect terminology.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhfxrBd9TAuzLZqSs7eu2_bxNeBow12yf54kiM1ow4pJsvDknFxfNQ7FpHPKp2MrkVe1_zpl2Kij4EdeL0ZpaUUIatS4mkEF8lTjFDH10X5Mwe7zdGdZN8bFLYAr_XBziKpDuFNd-IC0wMAJJ4dvVwaXzZlTHQOxLWmQQ3zbP6mcvFAUtYltLob3wr4m_fa/s1600/trust-attention-to-boost-quality-01.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="457" data-original-width="1600" height="114" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhfxrBd9TAuzLZqSs7eu2_bxNeBow12yf54kiM1ow4pJsvDknFxfNQ7FpHPKp2MrkVe1_zpl2Kij4EdeL0ZpaUUIatS4mkEF8lTjFDH10X5Mwe7zdGdZN8bFLYAr_XBziKpDuFNd-IC0wMAJJ4dvVwaXzZlTHQOxLWmQQ3zbP6mcvFAUtYltLob3wr4m_fa/w400-h114/trust-attention-to-boost-quality-01.png" width="400" /></a></div><br /><h3 style="text-align: left;">The Emergence of LLM-Based Translation Models</h3><p>In
the summer of 2023, we are at an interesting junction in the
development of AI-based language translation technology, where we now
see that Large Language Models (LLMs) are also an emerging technological
approach to having machines perform the language translation task. LLMs
are particularly impressive in handling idioms and enhancing the
fluency of machine translations.</p><p><strong>However, at this point,
there are still serious latency, high training, and inference costs, and
most importantly trustworthiness issues with the output produced by
Generative AI models like GPT-4.</strong> These issues will need to be addressed for Gen AI models to be viable in production-use translation settings. There is also the issue of poor performance in low-resource languages and a bias toward better performance with systems that translate into English.</p><p>The
AI product team at Translated continues to research and investigate the
possibilities for continued improvement of pure NMT models, hybrid NMT
and Gen AI models, as well as pure Gen AI models. <strong>Special
consideration is given to ensure that any major improvements made in
existing NMT model technology can also be leveraged in the future with
potentially production-use capable Gen AI translation models.</strong></p><p>AI
systems are trained on large datasets found on the internet, data that
can be of varied quality and reliability. If the data used for training
is biased or of poor quality, it can lead to biased or unreliable AI
outputs, and we have seen that one of the biggest obstacles to the
widespread use of Gen AI in mission-critical applications has been the
high levels of problematic and fluent, but untrustworthy output.</p><p>Better
data validation and verification can indeed improve the trustworthiness
of AI output. Data validation involves ensuring that the data used to
train and evaluate AI models is accurate, consistent, and representative
of the real-world scenarios the AI system will encounter. This can be
done through data cleaning, data preprocessing techniques, and careful
selection of training data.</p><p><br /></p><h1 style="text-align: left;"><span style="color: #2b00fe;">The Importance of Data Quality</span></h1><p>With this in mind, ModernMT Version 7, introduces a significant upgrade to its core adaptive machine translation (MT) system. <strong>This
new version introduces Trust Attention, a novel technique inspired by
how human researchers prioritize information from trusted sources</strong> and the V 7 model preferentially uses identified trustworthy data both in training and inference. </p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">This
innovation is the first of a long-term thematic effort focused on
improving data quality being undertaken at Translated, to ensure that
data quality and trustworthiness is a pervasive and comprehensive
attribute of all new translation AI initiatives.</span></strong></h3><p>Translated
has realized from a large number of independent evaluations and
internal testing over the years, that this focus on data quality enables
ModernMT to compare favorably in quality performance evaluations to
many other better-funded public generic MT engines produced by Google,
Microsoft, and others. </p><p>They have developed a robust data
governance framework to define data quality standards, processes, and
roles over the last decade. This helps create a culture of data quality
and ensures that data management practices are aligned with
organizational efficiency goals and technology improvements. </p><p>This
culture, together with close long-term collaboration with translators
ensures that ongoing data replenishment is of the highest quality and
systematically identifies and removes lower-quality data. Finally, <strong>regularly
measuring and monitoring data quality metrics helps to identify and
address potential issues before they impact AI performance. </strong></p><h3 style="text-align: left;"><span style="color: #2b00fe;">Trust
Attention is possible because of the long-term investment in developing
a data-quality culture that produces the right data to feed innovation
in new AI technologies.</span></h3><p></p><p>While it is common practice in the
industry to use automated algorithm-driven methods to drive data
validation and verification practices, Translated’s 20 years of
experience working with human translators show that human-verified data
is the most trustworthy data available to drive the learning of language
AI models. </p><p><strong>This human-verified data foundation is
precisely the most influential driver of preferential learning in the
ModernMT Version 7 models.</strong> Automated cleaning and verification
are valid ways to enhance data quality in machine learning applications,
but 10 years of experience show that human-verified data provide a
performance edge that is not easily matched by large-scale automated
cleaning and verification methods.</p><p><strong>Human quality assessments made comparing ModernMT V6 output versus V7 output show that the use of Trust Attention</strong> <strong>improves translation quality by as much as 42% of the time based on human evaluations.</strong> It is interesting to note that many high-resource languages like
Spanish, Chinese, and Italian also saw major improvements near the 30%
range in human evaluations. </p><p>Human evaluations and judgments are
corroborated by concurrent BLEU and COMET score measurements which are
also used to ensure that conclusions being drawn by introducing new
technology are accurate and trustworthy.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgANsn3ACUFgzS6kIGlbwayAnRLItO9LhnsaFTFGsfB8KgkI1q5Ph0mhZZpLGGSKkLlP8BsTFpmssOpNM_2OAGoYsubHeZCAkeb13qUBwB566HIbmpnlIwwW-ik6234vhflKXFWcfgz1f1S_4TepGBMGPpOR509TX-F0sFOZ189J3ZUq24NxaaLvTwtgv5n/s1600/trust-attention-to-boost-quality.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="557" data-original-width="1600" height="139" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgANsn3ACUFgzS6kIGlbwayAnRLItO9LhnsaFTFGsfB8KgkI1q5Ph0mhZZpLGGSKkLlP8BsTFpmssOpNM_2OAGoYsubHeZCAkeb13qUBwB566HIbmpnlIwwW-ik6234vhflKXFWcfgz1f1S_4TepGBMGPpOR509TX-F0sFOZ189J3ZUq24NxaaLvTwtgv5n/w400-h139/trust-attention-to-boost-quality.png" width="400" /></a></div><br /><p>The following is a sample of MT output from the ModernMT V7 system
compared to the previous V6. Three independent professional reviewers
were shown two randomized samples of a translation of the same source
segment and asked to judge if one was better, no different, or worse. The chart above
shows how often the V7 translation was preferred by a majority of the reviewers by language.</p><p>Examples below show sample sentences from English to Brazilian Portuguese and Simplified Chinese.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFVb8ko0oXcrxewhKb0-3-P8-LgmtGpQdm5wNAi7psE_a5DJ6P2TsT3IfK-ym5WnZfyghLTi_A8LspdWFtYkMXGakTc06xKAF97tk-feBFefFU-8n9rKb_csuj1x0lBslhM-QOwA6MrFYWA31GJiUV6YzZ-GJ7pRhiu3xVkD7sc16VY8-RO9gEG430oEZs/s960/B-PT.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="225" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhFVb8ko0oXcrxewhKb0-3-P8-LgmtGpQdm5wNAi7psE_a5DJ6P2TsT3IfK-ym5WnZfyghLTi_A8LspdWFtYkMXGakTc06xKAF97tk-feBFefFU-8n9rKb_csuj1x0lBslhM-QOwA6MrFYWA31GJiUV6YzZ-GJ7pRhiu3xVkD7sc16VY8-RO9gEG430oEZs/w400-h225/B-PT.png" width="400" /></a></div><br /><h1 style="text-align: left;"><span style="color: #2b00fe;">“If 80 percent of our work is data preparation, then ensuring data quality is the important work of a machine learning team.”</span></h1><p>Andrew Ng, Professor of AI at Standford University and founder of <a href="https://www.deeplearning.ai/the-batch/issue-84/?ref=blog.modernmt.com"><em>DeepLearning.AI</em></a></p><p><br /></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiTPsvscBYKCY5yFFizlfxkRpzme-WxXAqWzV51up7CQ2SaeNe27wWx6tNCQamVH3ejIX1-iEL7ACE5-O_DQZcx1Nxk5fJUOh6IQItY2N6XmVFmXggJ_e_AVc-T9b9Nz-yzhXIKJ7I1M9Yf1VzHI-zdY26vT7PmqDwzH4yrAYDsRSNV9zftwOvenJp3XezA/s960/ZH-V7.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="225" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiTPsvscBYKCY5yFFizlfxkRpzme-WxXAqWzV51up7CQ2SaeNe27wWx6tNCQamVH3ejIX1-iEL7ACE5-O_DQZcx1Nxk5fJUOh6IQItY2N6XmVFmXggJ_e_AVc-T9b9Nz-yzhXIKJ7I1M9Yf1VzHI-zdY26vT7PmqDwzH4yrAYDsRSNV9zftwOvenJp3XezA/w400-h225/ZH-V7.png" width="400" /></a></div><br /><h1 style="text-align: left;">How is Trust Attention Different?</h1><p>“Garbage
in, garbage out” (GIGO) is a concept in computing and artificial
intelligence (AI) that highlights the importance of input data quality.
It means that if the input data to a system, such as an AI model or
algorithm, is of poor quality, inaccurate, or irrelevant, the system’s
output will also be of poor quality, inaccurate, or irrelevant.</p><p>This
concept is particularly significant in the context of AI models which
use machine learning and deep learning models, and rely heavily on the
data used for training and validation. If the training data is biased,
incomplete, or contains errors, the AI model will likely produce
unreliable or biased results.</p><h2 id="all-data-is-not-equally-important"><br /></h2><h3 style="text-align: left;">All Data Is Not Equally Important</h3><p>Traditional
MT systems generally are not able to distinguish between trustworthy
data and lower-quality training material during the training process,
and typically all the data has equal weight. Thus, high-quality data and
high-volume noisy data can have essentially the same amount of impact
on how a translation model will perform. </p><h1 style="text-align: left;"><span style="color: #2b00fe;">Trust Attention allows an engine to prioritize more trustworthy data and have this data influence ongoing model behavior more heavily.</span></h1><p></p><p>ModernMT now uses a <strong>first-of-its-kind weighting system to enable primary learning from high-quality, trusted, and verified data </strong>– translations performed and/or reviewed by professional translators – over unverified data that is acquired from the Web.</p><p>As with adaptive MT, Translated looked to established human practices to develop this new technique. <strong>In
any serious research, humans collect and sift through multiple
information sources to identify and assign preferential status to the
most trustworthy and reliable data sources</strong>. </p><p><strong>ModernMT
V7 similarly identifies the most valuable training data and prioritizes
its learning based on certified and verified data by modeling this
human behavior. </strong>This certification and verification is not an
automated machine-led process, rather it is an expert human validation
that raises the trustworthiness of the data.</p><h1 style="text-align: left;"><span style="color: #2b00fe;"><strong>This focus on
prioritizing the use of trusted, verified data is a major step forward
in the development of enterprise-focused MT technology</strong>. </span></h1><p>The
efforts made to identify and build repositories of high-quality data
will also be useful in the future if there is indeed a shift to Gen
AI-based language translation models.</p><p>Today, there is considerable
discussion regarding the application of large language models in
translation. While the traditional NMT models seem to perform much
better on the accuracy dimension, though they can be less fluent than
humans, LLMs tend to emphasize and often win on fluency, even though
these models often produce misleading output due to hallucinations
(generative fabrication). </p><p><strong>Trust Attention methodology
deployed in LLMs, will also enhance the accuracy of generative models,
reducing the chances of random fabrication and confabulation errors.
This could set the stage for an emerging era of new machine translation
methodologies, one that combines the accuracy of dynamic adaptive NMT
with the fluency of Gen AI models.</strong></p><p>ModernMT Version 7
also introduces a data-cleaning AI that minimizes the likelihood of
hallucinations, making it valuable for companies seeking greater
accuracy in high-volume automated translation use cases, and is also
useful for translators integrating MT into their workflow.</p><p>John
Tinsley, VP of AI Solutions at Translated, added, "We are confident that
these new data validation and verification techniques can also improve
accuracy in generative AI systems, paving the way for the next
generation of machine translation."</p><p>The introduction of this new approach is a major step forward for companies seeking <strong>greater accuracy in the translation of large volumes of content</strong> or requiring a <strong>high degree of customization</strong> of the MT engine, as well as for translators integrating MT into their workflow.</p><h3 style="text-align: left;"><span style="color: #2b00fe;">The
combined impact of these multiple innovations provides global
enterprises with a superior platform to rapidly transform generic
engines into highly tuned enterprise-specific translation engines.</span></h3>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0tag:blogger.com,1999:blog-6748877443699290050.post-89122137645766694262023-12-05T11:07:00.000-08:002023-12-05T11:07:57.415-08:00The English-Centric Bias of Large Language Models<p> The internet is the primary source of information, economic opportunity, and community for many worldwide. However, <strong>the
automated systems that increasingly mediate our interactions online —
such as chatbots, content moderation systems, and search engines — are
primarily designed for and work far more effectively in English than in
the world’s other 7,000 languages</strong></p><p>It is clear to anyone
who works with LLMs and multilingual models, that there are now many
powerful and impressive LLM models available for generating natural and
fluent texts in English. While there has been substantial hype around
the capabilities and actual potential value of a wide range of
applications and use cases, the benefits have been most pronounced for
English-speaking users.</p><p>It is also now increasingly being
understood that achieving the same level of quality and performance for
other languages, even the ones that are widely spoken, is not an easy
task. <strong>AI chatbots are less fluent in languages other than
English and are thus threatening to amplify the existing language bias
in global commerce, knowledge access, basic internet research, and
innovation.</strong></p><p>In the past, it has been difficult to develop
AI systems — and huge language models in particular — in languages other
than English because of what is known as the<strong> resourcedness gap.</strong></p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">The
resourcedness gap describes the asymmetry in the availability of
high-quality digitized text that can serve as training data for a large
language model and generative AI solutions in general.</span></strong></h3><p></p><p>English
is an extremely highly resourced language, whereas other languages,
including those used predominantly in the Global South,<a href="https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00447/109285/Quality-at-a-Glance-An-Audit-of-Web-Crawled?ref=blog.modernmt.com"> often have fewer examples of high-quality text</a> (if any at all) on which to train language models.</p><h3 style="text-align: left;"><strong><span style="color: #2b00fe; font-size: large;">English-speaking
users have a better user experience with generative AI than users who
speak other languages, and the current models will only amplify this
English bias further.</span></strong></h3><p><a href="https://www.springboard.com/blog/data-science/machine-learning-gpt-3-open-ai/?ref=blog.modernmt.com#:~:text=Although%20GPT%2D3's%20training%20data,on%20the%20language%20translation%20task.">It is estimated</a>
that although GPT-3's training data consists of > 90% English text
it did include some foreign language text, but not enough to ensure that
model performance across different languages is consistent. GPT-3 was
the foundation model used to build ChatGPT and though we do not know
what data was used in GPT-4 we can safely assume that no major sources
of non-English data have been acquired, primarily because it is not
easily available.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgokDmnNufyx4ibbQruRbml4oUEQFefN4osIkhpiv0HBSKV0zvLKUaVbAAlq5FZOxqyvlZeoJjvgHTrMWDC8_rmPiOUmKEo9xQp9eLTQF_NvVN-IpY_F8_KPD36cz8efAjhdl_3NXFX2evFOmQreya4y3oKkSuhCRT6gr0JV69vtGq_llsP3uLiGC9-gmWe/s2000/The-Problem-of-English-Dominated-Generative-AI_converted-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img alt="ategories of language resourcedness. Languages divided into different levels of resourcedness, according to labeled and unlabeled datasets available as of 2020" border="0" data-original-height="865" data-original-width="2000" height="173" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgokDmnNufyx4ibbQruRbml4oUEQFefN4osIkhpiv0HBSKV0zvLKUaVbAAlq5FZOxqyvlZeoJjvgHTrMWDC8_rmPiOUmKEo9xQp9eLTQF_NvVN-IpY_F8_KPD36cz8efAjhdl_3NXFX2evFOmQreya4y3oKkSuhCRT6gr0JV69vtGq_llsP3uLiGC9-gmWe/w400-h173/The-Problem-of-English-Dominated-Generative-AI_converted-1.jpg" width="400" /></a></div><strong><span style="font-size: x-small;">Source:<a href="https://cdt.org/wp-content/uploads/2023/05/non-en-content-analysis-primer-051223-1203.pdf?ref=blog.modernmt.com"><strong> Lost in Translation Large Language Models in Non-English Content Analysis</strong></a></span></strong><div><b><br /></b></div><div><p>Researchers like Pascale Fung and others have pointed out the
difficulty for many global customers because of the dominance of English
in eCommerce. <strong>It is much easier to get information about products in English in online marketplaces than it is in any other language.</strong></p><p>Fung,
director of the Center for AI Research at the Hong Kong University of
Science and Technology, who herself speaks seven languages, sees this
bias even in her research field. <strong>“If you don’t publish papers in
English, you’re not relevant,” she says. “Non-English speakers tend to
be punished professionally.”</strong></p><p>The<a href="https://www.sigmoid.com/blogs/gpt-3-all-you-need-to-know-about-the-ai-language-model/?ref=blog.modernmt.com"> following table</a> describes the source data for the training corpus of GPT-3 which is the data foundation for ChatGPT:</p><p></p><table border="0" cellpadding="0" cellspacing="0" class="MsoNormalTable" style="border-collapse: collapse; mso-padding-alt: 0in 0in 0in 0in; mso-yfti-tbllook: 1184;">
<tbody><tr>
<td style="background: #E2CC6F; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:black="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Datasets<o:p></o:p></span></span></strong></p>
</td>
<td style="background: #E2CC6F; border-left: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:black="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Quantity (Tokens)<o:p></o:p></span></span></strong></p>
</td>
<td style="background: #E2CC6F; border-left: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:black="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Weight in Training Mix<o:p></o:p></span></span></strong></p>
</td>
<td style="background: #E2CC6F; border-left: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:black="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Epochs elapsed when
training for 300 BN tokens<o:p></o:p></span></span></strong></p>
</td>
</tr>
<tr style="height: 42.15pt; mso-yfti-irow: 1;">
<td style="border-top: none; border: solid gainsboro 1.0pt; height: 42.15pt; mso-border-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Common Crawl
(filtered)<o:p></o:p></span></span></strong></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; height: 42.15pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">410 BN<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; height: 42.15pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">60%<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; height: 42.15pt; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" valign="top" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">0.44<o:p></o:p></span></span></p>
</td>
</tr>
<tr>
<td style="border-top: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">WebText2<o:p></o:p></span></span></strong></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">19 BN<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">22%<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" valign="top" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">2.90<o:p></o:p></span></span></p>
</td>
</tr>
<tr>
<td style="border-top: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Books1<o:p></o:p></span></span></strong></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">12 BN<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">8%<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" valign="top" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">1.90<o:p></o:p></span></span></p>
</td>
</tr>
<tr>
<td style="border-top: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Books2<o:p></o:p></span></span></strong></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">55 BN<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">8%<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" valign="top" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">0.43<o:p></o:p></span></span></p>
</td>
</tr>
<tr>
<td style="border-top: none; border: solid gainsboro 1.0pt; mso-border-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><strong><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">Wikipedia<o:p></o:p></span></span></strong></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">3 BN<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt;" valign="top">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">3%<o:p></o:p></span></span></p>
</td>
<td style="border-bottom: solid gainsboro 1.0pt; border-left: none; border-right: solid gainsboro 1.0pt; border-top: none; mso-border-alt: solid gainsboro .75pt; mso-border-left-alt: solid gainsboro .75pt; mso-border-top-alt: solid gainsboro .75pt; padding: 7.5pt 15.0pt 7.5pt 15.0pt; width: 172.9pt;" valign="top" width="231">
<p align="center" class="MsoNormal" style="line-height: normal; margin-bottom: 0in; text-align: center;"><span color:="" mso-bidi-font-family:calibri="" mso-bidi-theme-font:minor-latin="" mso-font-kerning:0pt="" mso-ligatures:none="" new="" roman="" times=""><span style="font-size: x-small;">3.40</span><span style="font-size: 12pt;"><o:p></o:p></span></span></p>
</td>
</tr>
</tbody></table><p>Understanding what data has been used to train GPT-3 is useful.<a href="https://jilltxt.net/right-now-chatgpt-is-multilingual-but-monocultural-but-its-learning-your-values/?ref=blog.modernmt.com"> This overview</a> provides some valuable details that also help us understand the English bias and US-centric perspective that these models have.</p><p>Fung and others are part of a global community of AI researchers<a href="https://arxiv.org/pdf/2302.04023.pdf?ref=blog.modernmt.com"> testing the language skills of ChatGPT</a>
and its rival chatbots and sounding the alarm about providing evidence
that they are significantly less capable in languages other than
English.</p><ul><li><strong>ChatGPT still lacks the ability to understand and generate sentences in low-resource languages.</strong> The performance disparity in low-resource languages limits the diversity and inclusivity of NLP.</li><li><strong>ChatGPT also lacks the ability to translate sentences in non-Latin script languages</strong>, despite the languages being considered high-resource.</li></ul><p>“One
of my biggest concerns is we’re going to exacerbate the bias for
English and English speakers,” says Thien Huu Nguyen, a University of
Oregon computer scientist who is also a leading researcher raising
awareness about<a href="https://arxiv.org/pdf/2304.05613.pdf?ref=blog.modernmt.com"> the often impoverished experience non-English speakers routinely experience</a> with generative AI. Nguyen specifically points out:</p><p><strong>ChatGPT’s
performance is generally better for English than for other languages,
especially for higher-level tasks that require more complex reasoning
abilities </strong>(e.g., named entity recognition, question answering,
common sense reasoning, and summarization). The performance differences
can be substantial for some tasks and lower-resource languages.</p><ul><li><strong>ChatGPT can perform better with English prompts</strong> even though the task and input texts are intended for other languages.</li><li><strong>ChatGPT<a href="https://arxiv.org/pdf/2304.05613.pdf?ref=blog.modernmt.com"><strong> performed substantially</strong></a> worse at answering factual questions or summarizing complex text in non-English languages</strong> and was more likely to fabricate information.</li></ul><p>
</p><p>The research tends to point clearly to the English bias of the most popular LLMs and state: The AI systems are good at<a href="https://arxiv.org/pdf/2304.04675.pdf?ref=blog.modernmt.com"> translating other languages into English</a>, but they struggle with rewriting English into other languages—especially for languages like Korean, with non-Latin scripts.</p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">“51.3%
of pages are hosted in the United States. The countries with the
estimated 2nd, 3rd, and 4th largest English-speaking populations—India,
Pakistan, Nigeria, and The Philippines—have only 3.4%, 0.06%, 0.03%,
0.1% the URLs of the United States, despite having many tens of millions
of English speakers.”</span></strong></h3><div style="text-align: left;"><span style="font-size: x-small;">(<a href="http://arxiv.org/abs/2104.08758.?ref=blog.modernmt.com"><strong>Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus</strong></a>, 2021, p. 4)</span></div><p>The
chart below displays a deeper dive into the linguistic makeup of the
Common Crawl data by the Common Sense Advisory research team.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgvwX_8yBUthWrZlpExxG8sTR9LDajBVhhvu_xP964h56FOZ3iAiF0uZIesNCpjaso3U0vugKBRj-k8Ea7NoesZAbMTOK795VAvGOuSPubH4ajbUhqor3TmgIQNMK0HeO_lEqYAT5TN45GLhzkrlcYvqrHj8p9RvXxCgc5O94_QCSRvmFX2wR9EZVua0mnI/s1600/CSA-GAI-Data-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="840" data-original-width="1600" height="210" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgvwX_8yBUthWrZlpExxG8sTR9LDajBVhhvu_xP964h56FOZ3iAiF0uZIesNCpjaso3U0vugKBRj-k8Ea7NoesZAbMTOK795VAvGOuSPubH4ajbUhqor3TmgIQNMK0HeO_lEqYAT5TN45GLhzkrlcYvqrHj8p9RvXxCgc5O94_QCSRvmFX2wR9EZVua0mnI/w400-h210/CSA-GAI-Data-1.png" width="400" /></a></div><br /><p>Recently though, researchers and technology companies have attempted
to extend the capabilities of large language models into languages other
than English by building what are called multilingual language models.
Instead of being trained on text from only one language, multilingual
language models are trained on text from dozens or hundreds of languages
at once.</p><p>Researchers posit that multilingual language models can
infer connections between languages, allowing them to apply word
associations and underlying grammatical rules learned from languages
with more text data available to train on (in particular English) to
those with less.</p><p>Languages vary widely in resourcedness, or the
volume, quality, and diversity of text data they have available to train
language models on. <strong>English is the highest-resourced language by multiple orders of magnitude,</strong> but Spanish, Chinese, German, and a handful of other languages have sufficiently high resources to build language models.</p><p>However, they are still expected to be lower in quality than English language
models. Medium resource languages, with fewer but still high-quality
data sets, such as Russian, Hebrew, and Vietnamese, and low resource
languages, with almost no training data sets, such as Amharic, Cherokee,
and Haitian Creole, have too little text for training large language
models</p><p>However, there are many challenges and complexities
involved in developing multilingual and multicultural LLMs that can
cater to the diverse needs and preferences of different communities.
Multilingual language models are still usually trained
disproportionately on English language text and thus end up transferring
values and assumptions encoded in English into other language contexts
where they may not belong.</p><h3 style="text-align: left;"><span style="color: #2b00fe; font-size: large;">Most
remedial approaches to address the English bias rely on the acquisition
of large amounts of non-English data to be added to the core training
data to reduce the English bias in current LLMs; data which is not
easily found or often non-existent. Certainly not at the scale, volume,
and diversity that English training data exists.</span></h3><p>English is the
closest thing there is to a global lingua franca. It is the dominant
language in science, popular culture, higher education, international
politics, and global capitalism; it has the<a href="https://www.ethnologue.com/?ref=blog.modernmt.com"> most total speakers</a> and the third-most first-language speakers.</p><p><strong>The
bias in the NLP research community is evident in the chart below. ACL
papers are more likely to be published in English than any other
language by a factor of 11X to 80X!</strong></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjSVBR1c1osMYC4_N5tRPbP3-APqBkHwjs2-31PZLZm6EPIadfyECAiTDMsLOTGVZwtg1rNeWN0mqDwntyaTfa9nc2N7z0hgGWmfaX6iPqwclhk1WRdfe4ZIB8kdtUk5puruIIBbnD0lUNr907a-ciitjicE0uQiD_tWdZJPYaYpk8xODh2PWiwmqMGhDg3/s1600/The-Problem-of-English-Dominated-Generative-AI_converted2-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="848" data-original-width="1600" height="170" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjSVBR1c1osMYC4_N5tRPbP3-APqBkHwjs2-31PZLZm6EPIadfyECAiTDMsLOTGVZwtg1rNeWN0mqDwntyaTfa9nc2N7z0hgGWmfaX6iPqwclhk1WRdfe4ZIB8kdtUk5puruIIBbnD0lUNr907a-ciitjicE0uQiD_tWdZJPYaYpk8xODh2PWiwmqMGhDg3/s320/The-Problem-of-English-Dominated-Generative-AI_converted2-1.jpg" width="320" /></a></div><figure class="kg-card kg-image-card kg-width-wide kg-card-hascaption"><figcaption><b><strong style="white-space-collapse: preserve;">Languages mentioned in paper abstracts. Top most mentioned languages in abstracts of papers published by the Association for Computational Linguistics, May 2022-January 2023.</strong></b></figcaption></figure><p>Recent<a href="https://www.wired.com/story/spooked-by-chatgpt-us-lawmakers-want-to-create-an-ai-regulator/?ref=blog.modernmt.com"> US congressional hearings also focused on this language-bias problem</a>
when Senator Alex Padilla (a native Spanish speaker) of California
questioned the CEO of OpenAI about improving the experience for the
growing population of non-English users even in the US and said: <strong>“These
new technologies hold great promise for access to information,
education, and enhanced communication, and we must ensure that language
doesn’t become a barrier to these benefits.”</strong></p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">However,
the fact remains, and OpenAI clearly states that the majority of the
underlying training data used to power ChatGPT (and most other LLMs)
came from English and that the company’s efforts to fine-tune and study
the performance of the model is primarily focused on English </span><span style="color: red;">“with a
US-centric point of view.”</span></strong></h3><p>This also results in the models performing better on tasks that involve
going from Language X to English than on tasks that involve going from
English to Language X. Because of the data scarcity and substantial
costs involved in correcting this it is not likely to change soon.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiIu7sTDzBbzUt1zGtoBYTX90qENF5N_u2ynftzh369vwDrmMMTTV5ZENuj54YkPXwB1o_RHeafOONOCo_ndnJOiez15zaMNgFz_egxgpyGxgfxP5JIrqoPldprgOsYUmU-2oondUNHe2iZzcpUxR-C5o0hMiR6T9JhInHGIqo-4vuPBGE7_dpZv2e4Team/s1600/The-Problem-of-English-Dominated-Generative-AI_converted3-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="692" data-original-width="1600" height="173" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiIu7sTDzBbzUt1zGtoBYTX90qENF5N_u2ynftzh369vwDrmMMTTV5ZENuj54YkPXwB1o_RHeafOONOCo_ndnJOiez15zaMNgFz_egxgpyGxgfxP5JIrqoPldprgOsYUmU-2oondUNHe2iZzcpUxR-C5o0hMiR6T9JhInHGIqo-4vuPBGE7_dpZv2e4Team/w400-h173/The-Problem-of-English-Dominated-Generative-AI_converted3-1.jpg" width="400" /></a></div><p>Because the training text data sets used to train GPT models also
have some other languages mixed in, the generative AI models do pick up
some capability in other languages. However, their knowledge is not
necessarily comprehensive or complete enough, and in a development
approach that implicitly assumes that scale is all you need, most
languages simply do not have enough scale in training data to perform at
the same levels as English.</p><p>This is likely to change over time to
some extent, and already the Google PaLM model claims to be able to
handle more languages, but early versions show only very small
incremental improvements in a very few select languages.</p><h3 style="text-align: left;"><strong><span style="color: red; font-size: large;">Each
new language that is "properly supported" will require a separate set
of guardrails and controls to minimize problematic model behavior.</span></strong></h3><p></p><p>Thus,
beyond the monumental task of finding massive amounts of non-English
text and re-training the base generative AI model from scratch,
researchers are also trying other approaches e.g., creating new data
sets of non-English text to try to accelerate the development of truly
multilingual models, or by generating synthetic data by using what is
available in high resource languages like English or Chinese, <strong>which are both less effective than simply having the adequate data volume in the low-resource language in the first place.</strong></p><p>Nguyen
and other researchers say they would also like to see AI developers pay
more attention to the data sets they feed into their models and better
understand how that affects each step in the building process, not just
the final results. So far, the data and which languages end up in models
has been a "random process," Nguyen says.</p><p><strong>So when you
make a prompt request in English, it draws primarily from all the
English language data it has. When you make a request in traditional
Chinese, it draws primarily from the Chinese language data it has. How
and to what extent these two piles of data inform one another or the
resulting outcome is not clear, but at present, experiments show that
they at least are quite independent.</strong></p><p>The training data
for these models were collected through long-term web crawling
initiatives, and a lot of it was pretty random. More rigorous controls
to reach certain thresholds of content for each language -as Google
tried to do with PaLM- could improve the quality of non-English output.
It is also possible that more carefully collected and curated data that
is better balanced linguistically could improve performance across more
languages.</p><h1 style="text-align: left;"><br />The T-LM (Translated Language Model) Offering</h1><p></p><p><strong>The fundamental data acquisition and limited and sub-optimal accessibility problems described above could take years to resolve</strong>.
Thus, Translated Srl is introducing a way to address the needs of a
larger global population interested in using GPT-4 for content creation,
content analysis, basic research, and content refinement in their
preferred language.</p><p>The following chart shows the improved
performance available with T-LM across several languages. Users can
expect the performance improvements to continue to increase and improve
as they provide corrective feedback daily.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhatf4l1kkbE659IwLJILcvaPUbP6Cv4Gwscf9Yn3lCpuE7Pr__ZNahZpNDfEpKvRKfTn7eiSuDOpapUj2hfHayvSe-d76uVvLS3yyobtpSCSsG9jV9Vfbj9cvQHjyqxqu7SP7pckVm5SJr_Xp8NfO3Vzwn11lKQZo51Tdt-K3Z6UnRMYV51khjGOntevBm/s1600/The-Problem-of-English-Dominated-Generative-AI_converted4-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="692" data-original-width="1600" height="173" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhatf4l1kkbE659IwLJILcvaPUbP6Cv4Gwscf9Yn3lCpuE7Pr__ZNahZpNDfEpKvRKfTn7eiSuDOpapUj2hfHayvSe-d76uVvLS3yyobtpSCSsG9jV9Vfbj9cvQHjyqxqu7SP7pckVm5SJr_Xp8NfO3Vzwn11lKQZo51Tdt-K3Z6UnRMYV51khjGOntevBm/w400-h173/The-Problem-of-English-Dominated-Generative-AI_converted4-1.jpg" width="400" /></a></div><p>Combining the power of the state-of-the-art adaptive machine
translation technology with OpenAI's latest language model will result
in empowering users<a href="https://blog.modernmt.com/modernmt-significantly-expands-language-coverage/"> across 200 languages</a> to engage and explore the capabilities of GPT-4 in a preferred non-English language and achieve superior performance.</p><p>T-LM will help unlock the full potential of GPT-4 for businesses around the world. It provides companies with a <strong>cost-effective solution to create and restructure content and do basic content research in 200 languages</strong>, bridging the performance gap between GPT-4 in English and non-English languages.</p><p>A detailed overview of<a href="https://blog.modernmt.com/modernmt-significantly-expands-language-coverage/"> the 200 specific languages and their importance in the changing global dynamics is described here</a>.</p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">Many
users have documented and reported sub-optimal performance when
searching with Bing Chat when they query in Spanish rather than English.</span></strong></h3><p></p><p>In
a separate dialog, when queried in English, Bing Chat correctly
identified Thailand as the rumored location for the next set of the TV
show<a href="https://www.wired.com/story/white-lotus-scene-epitomized-2022/?ref=blog.modernmt.com"> <em>White Lotus</em></a>,
but provided “somewhere in Asia” when the query was translated to
Spanish, says Solis, who runs a consultancy called Orainti that helps
websites increase visits from search engines.</p><p>Other discussions point out that<a href="https://www.reddit.com/r/ChatGPT/comments/134i1rp/does_chat_gpt_perform_much_worse_in_other/?onetap_auto=true&ref=blog.modernmt.com"> ChatGPT performs sub-optimally in most languages other than English.</a> Techcrunch also ran some tests to demonstrate that<a href="https://techcrunch.com/2023/04/26/why-chatgpt-lies-in-some-languages-more-than-others/?ref=blog.modernmt.com"> ChatGPT has lesser performance in non-English</a> languages.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj9PhXBZuQ0qdtgPRwweMU3zupE6LYzmGQHvP_Hk1c4-nzeycfXY8URMwiIsxSEicMH2xWfhcStYZ2mj6QslFBRooZ_up_6WIoRQxcVYgHs-huHSOF9cp2d8dy5fOaCI4iIbVyv5uj-4bbZxfhDW6p6s0Lt0xAMuItVgpgN9ZOkPO8IB3XzrAN7EkiZH6e_/s953/Twitter%20msg.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="925" data-original-width="953" height="389" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj9PhXBZuQ0qdtgPRwweMU3zupE6LYzmGQHvP_Hk1c4-nzeycfXY8URMwiIsxSEicMH2xWfhcStYZ2mj6QslFBRooZ_up_6WIoRQxcVYgHs-huHSOF9cp2d8dy5fOaCI4iIbVyv5uj-4bbZxfhDW6p6s0Lt0xAMuItVgpgN9ZOkPO8IB3XzrAN7EkiZH6e_/w400-h389/Twitter%20msg.png" width="400" /></a></div><p>Additionally, using GPT-4 in non-English languages can <strong>cost up to 15 times more</strong>
(see the charts below). Research has shown that speakers of certain
languages may be overcharged for language models while obtaining poorer
results, indicating that tokenization may play a role in both the cost
and effectiveness of language models.<a href="https://www.researchgate.net/figure/Estimated-cost-per-language-family-script-relative-to-English-The-language-families-are_fig2_370981335?ref=blog.modernmt.com"> This study shows the difference in cost by language family</a> which can be significantly higher than English.</p><p><a href="https://blog.yenniejun.com/p/all-languages-are-not-created-tokenized?ref=blog.modernmt.com">Independent researchers point out how the same prompt varies across</a>
languages and that some languages consistently have a higher token
count. Languages such as Hindi and Bengali (which together over 800
million people speak) resulted in a median token length of about 5 times
that of English. The ratio is 9 times that of English for Armenian and
over 10 times that of English for Burmese. <span style="color: #2b00fe;">In other words, <strong>to express the same prompt or sentiment, some languages require up to 10 times more tokens</strong>.</span></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi-mmkl8xcJ9jqHOyTLFlCSh6Qnly9Auw8fiw4tTiIDO2bVhJ8oaQ5TtMKWnyW-cgtNuoJxIsQ5frsDsgjT7o8cInLx-NhwBWEkzsubY0nypQu_lh5yLreIOf45ICzFToUE_hUbfXDe_zutfrHjVCdWHX83R3mcECwdFslO0BgnjhFE-ady-Gi6aAni-yXC/s1009/Tokens-in-LLM-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="450" data-original-width="1009" height="179" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi-mmkl8xcJ9jqHOyTLFlCSh6Qnly9Auw8fiw4tTiIDO2bVhJ8oaQ5TtMKWnyW-cgtNuoJxIsQ5frsDsgjT7o8cInLx-NhwBWEkzsubY0nypQu_lh5yLreIOf45ICzFToUE_hUbfXDe_zutfrHjVCdWHX83R3mcECwdFslO0BgnjhFE-ady-Gi6aAni-yXC/w400-h179/Tokens-in-LLM-1.png" width="400" /></a></div><div class="separator" style="clear: both; text-align: center;"><figcaption><span style="font-size: xx-small;"><b><strong style="white-space-collapse: preserve;">Source:</strong></b><a href="https://blog.yenniejun.com/p/all-languages-are-not-created-tokenized?ref=blog.modernmt.com"><b><strong style="white-space-collapse: preserve;"> All languages are NOT created (tokenized) equal</strong></b></a></span></figcaption><figcaption><br /></figcaption><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi1494jXzAB5Gb2Hu6bR47NfLk9-6yYeWE6AwGRyr-KTiuF4MI8U6t_omFYyqNQm0OsEDDQl1tTR5zlRQKzOBjj743Rt0kE84B5PCpVQH9We_795fygVJjQFgkYc6558dQJ0S2TARS4DgkeD2ARHWqhljDF76htXVJshaiO5OJrJTzb8Pn57XCmcHZZyI7K/s862/Median-Token-Length-vs-English-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="436" data-original-width="862" height="203" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi1494jXzAB5Gb2Hu6bR47NfLk9-6yYeWE6AwGRyr-KTiuF4MI8U6t_omFYyqNQm0OsEDDQl1tTR5zlRQKzOBjj743Rt0kE84B5PCpVQH9We_795fygVJjQFgkYc6558dQJ0S2TARS4DgkeD2ARHWqhljDF76htXVJshaiO5OJrJTzb8Pn57XCmcHZZyI7K/w400-h203/Median-Token-Length-vs-English-1.png" width="400" /></a></div><br /><figcaption><figure class="kg-card kg-image-card kg-width-wide kg-card-hascaption"><figcaption><b><strong style="white-space-collapse: preserve;"><span style="font-size: x-small;">To express the same sentiment, some languages require up to 10 times more tokens</span></strong></b></figcaption></figure><p><br /></p><h3 style="text-align: left;">Implications of tokenization language disparity</h3><p style="text-align: left;">Overall, requiring more tokens (to tokenize the same message in a different language) means:</p><ul><li style="text-align: left;">Non-English users are limited in how much information they can put in the prompt (because the context window is fixed).</li><li style="text-align: left;">It is more costly as generally more tokens are needed for equivalent prompts.</li><li style="text-align: left;">It is slower and takes longer to run and often results in more fabrication and other errors. </li></ul><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">OpenAI’s
models are increasingly being used in countries where English is not
the dominant language. According to SimilarWeb.com, the United States
only accounted for 10% of the traffic sent to ChatGPT in Jan-March 2023.
India, Japan, Indonesia, and France all have large user populations
that are almost as large as the US user base. </span></strong></h3><p style="text-align: left;">Translated's
T-LM service integrates the company’s award-winning adaptive machine
translation (ModernMT) with GPT-4 to bring advanced generative AI
capabilities to every business in<a href="https://blog.modernmt.com/modernmt-significantly-expands-language-coverage/"> the languages spoken by 95% of the world's population.</a>
This approach also lowers the cost of using GPT-4 in languages other
than English, since the pricing model is based on text segmentation
(tokenization) that is optimized for English. By ensuring that all
prompts submitted to GPT-4 are in English the billing will be equivalent
to the more favorable and generally lower-cost English tokenization.
T-LM, instead, will always use the number of tokens in English for
billing.</p></figcaption></div></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhl46JZUgzhMXb80Kkksts6R4EhjSUVU37NscxRp-_-cdTaF9vkUURrDxyEAgVsLLRQgS3MKtUSlSse-1WwrWW-embiwY0gKwjLuhxrdDoIzM31lWZArYd3Rhr2C-nS6cpMzNLIB6Lfi2RBnYaS-7mzAKPvbSvCBkdPoKBGE24AwadTkHBS3Pr2TrBl_GmZ/s1600/gpt4-vs-tlm.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="848" data-original-width="1600" height="213" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhl46JZUgzhMXb80Kkksts6R4EhjSUVU37NscxRp-_-cdTaF9vkUURrDxyEAgVsLLRQgS3MKtUSlSse-1WwrWW-embiwY0gKwjLuhxrdDoIzM31lWZArYd3Rhr2C-nS6cpMzNLIB6Lfi2RBnYaS-7mzAKPvbSvCBkdPoKBGE24AwadTkHBS3Pr2TrBl_GmZ/w400-h213/gpt4-vs-tlm.png" width="400" /></a></div><p>The<a href="https://blog.modernmt.com/understanding-adaptive-machine-translation/"> Adaptive ModernMT technology</a>,
unlike most other MT technology available today can learn and improve
dynamically and continuously with ongoing corrective feedback daily.
Thus, users who work with T-LM can drive continuous improvements in
output produced from GPT-4 by providing corrective feedback on the
translations produced by T-LM. This is something that is not possible
with the most commonly used static MT systems where users would be
confined and limited to generic system performance.</p><p><strong>T-LM
addresses the performance disparity experienced by non-English users by
translating the initial prompt from the source language to English and
then back to the user's language using a specialized model that has been
optimized for the linguistic characteristics typically used in prompts.</strong></p><p>T-LM
combines GPT-4 with ModernMT, an adaptive machine translation engine,
to offer GPT-4 near English-level performance in 200 languages. </p><ul><li>T-LM
works by translating non-English prompts into English, executing them
using GPT-4, and translating the output back to the original language,
all using the ModernMT adaptive machine translation.</li><li>T-LM is available to enterprises via an API and to consumers through a ChatGPT plugin.</li></ul><h3 style="text-align: left;"><strong><span style="color: #800180;">The
result is a more uniform language model performance capability across
many languages and enhanced GPT-4 performance in non-English languages.</span></strong></h3><p></p><ul><li>Customers can optionally use their existing ModernMT keys to employ adaptive models within GPT-4.</li><li>An
indirect benefit of T-LM is that it has cost up to 15x lower than
GPT-4, thanks to a reduced number of tokens billed. GPT-4 counts
significantly more tokens in non-English languages. T-LM, instead, will
always use the number of tokens in English for billing</li></ul><p>Therefore,
Translated's integration with OpenAI enhances GPT-4's performance in
non-English languages by combining GPT-4 with the ModernMT adaptive
machine translation, resulting in a more uniform language model
capability across languages and lower costs.</p><h3 style="text-align: left;"><strong><span style="color: #800180;">Use
cases for T-LM include assisting global content creation teams in a
broad range of international commerce-related initiatives, allowing
companies from Indonesia, Africa, and various parts of India to make
their products visible in online eCommerce platforms to US and EU
customers, providing better multilingual customer support, making global
user-generated content visible and understandable in the customer’s
language.</span></strong></h3><p><br /><strong>T-LM can be used in many text analysis tasks needed in business settings</strong>, e.g., <strong>breaking down and explaining complicated topics</strong>,
outlining blog posts, sentiment analysis, personalized responses to
customers, summarization, creating email sales campaign material, or
suggesting answers to customer agents.</p><h3 style="text-align: left;"><strong><span style="color: #2b00fe;">T-LM
works together with GPT to create a wide range of written content or
augment existing content to give it a different intonation, by softening
or professionalizing the language, to improve content creation and
transformation automation while providing a fast and engaging user
experience. This is now possible to do in 200 languages that ModernMT
supports.</span></strong></h3><p></p><p>There are many ways GPT-4 can produce
‘draft’ text that meets the length and style desired, which can then be
reviewed by the user,”<a href="https://www.computerworld.com/article/3687614/how-enterprises-can-use-chatgpt-and-gpt-3.html?ref=blog.modernmt.com"> Gartner said in a report on how to use GPT-4.</a>
“Specific uses include drafts of marketing descriptions, letters of
recommendation, essays, manuals or instructions, training guides, social
media or news posts.”</p><p>T-LM will allow students around the world
to access knowledge content, and use GPT-4 as a research assistant and
access a much larger pool of information. In education, GPT-4 can be
used to create personalized learning experiences, as a tutor would. And,
in healthcare, chatbots and applications can provide simple language
descriptions of medical information and treatment recommendations.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhB5aFym2CFsAVRQcha5V5-IqCCcC_3dW60jjrFNSAOzisNiTlLb4hVp55IqT_sIMjyDcM7NQIY8yrXLSO2SzZus_5bc-bOGnxGhaNcwA_odDbNeDRawUcjD6lID_YmqQtXYaqskqk-Tn-ILMwz2dehxSI1tm76i12jxjHLuAgc_FSOLpEJFaJzC1DiySF_/s1549/mateTavola-disegno-1@3x-100-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1549" data-original-width="1000" height="400" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhB5aFym2CFsAVRQcha5V5-IqCCcC_3dW60jjrFNSAOzisNiTlLb4hVp55IqT_sIMjyDcM7NQIY8yrXLSO2SzZus_5bc-bOGnxGhaNcwA_odDbNeDRawUcjD6lID_YmqQtXYaqskqk-Tn-ILMwz2dehxSI1tm76i12jxjHLuAgc_FSOLpEJFaJzC1DiySF_/w259-h400/mateTavola-disegno-1@3x-100-1.jpg" width="259" /></a></div><br /><p><strong>T-LM will enhance the ability of large and SME businesses to
engage in new international business by assisting in basic
communication, and understanding, and providing more complete
documentation on business proposals using the strengths of both GPT-4
and T-LM working together.</strong></p><p>T-LM is available now through API. More information on the service can be found at<a href="https://translatedlabs.com/gpt?ref=blog.modernmt.com"> translatedlabs.com/gpt</a>.</p>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com2tag:blogger.com,1999:blog-6748877443699290050.post-33465115057106744482023-06-23T15:27:00.000-07:002023-06-23T15:27:42.544-07:00The MT languages that will matter most over the next 50 years<p> Translated recently announced that ModernMT now supports 200
languages, setting a new benchmark in the industry. No other commercial
MT service currently supports such an extensive range of languages. By
expanding its language coverage to a potential reach of 6.5 billion
people, Translated enhances the ability of enterprises to create
stronger connections with their users and customers worldwide, fostering
better communication and understanding of a larger global customer
base.</p><p>In the modern era, we are rapidly moving to a world where a
global enterprise needs to expand the scope and nature of its
communication and information sharing with the modern digital customer.
There is a relationship between content strategy, e-commerce, and MT, as
much of this new content that enhances the customer experience is
constantly changing, and there is great value in making it multilingual
to enable engagement with a broader global customer base.</p><p><strong>The modern digital-first customer demands and expects much more relevant information from every organization they interact with.</strong>
The lack of needed information can easily trigger a potential customer
to walk away from a brand and company that may otherwise be a very good
fit in terms of matching customer needs to available product offerings.
We know that today:</p><ul><li>The modern buyer and customer journey has many digital touchpoints.</li><li>Global-savvy
companies are increasingly moving to a business model focused on
customer needs, attempting to serve as much information as needed to
improve the global customer experience.</li><li><strong>Companies are
translating everything that might be useful to a customer, not just what
is mandated by local commercial regulations</strong>.</li></ul><p>So,
what are some of the things that are required of a globally focused
business to be successful with an ever-growing global customer base?</p><p>The list of actions recommended by globalization consultants on <strong>best practices for providing relevant information to the modern digital-first customer includes all of the following actions:</strong></p><ul><li><strong>Personalize</strong> communication and content to their interests and needs</li><li>Provide easy access to information through <strong>self-service portals</strong> and knowledge bases</li><li><strong>Utilize chatbots and AI technology</strong> to provide instant and accurate answers to common questions</li><li><strong>Collect customer feedback </strong>to continuously improve and adjust information and communication strategies</li><li><strong>Offer various communication channels</strong> for different preferences, such as email, phone, social media, and messaging apps.</li></ul><p><strong>A
global enterprise can immediately establish a comprehensive digital
presence in international markets if they solve the translation
challenge and make relevant content multilingual at scale. </strong>The
notion of digital-first applies both to the global enterprise and the
global customer. Digital-first also means that it is much easier for a
company to start engaging with a global customer base. A digital-first
globalization strategy allows the enterprise to expand rapidly and be
customer-centric across the world.</p><p><strong>Never has it been more
critical for a company to translate its content into many different
languages to quickly establish a relationship with global customers</strong>.
The sheer volume and broad scope of the task require that the
translation challenge needs to be handled in a way that is efficient,
streamlined, and scalable. </p><p>To achieve a comprehensive digital
presence, a global enterprise needs to focus on solving the translation
challenge and making all relevant content multilingual rather than
trying to minimize the translated content volume. </p><div style="text-align: left;"><span style="font-size: large;"><span style="color: #2b00fe;"><b>The modern era requirement in the digital age is to "translate everything".</b></span></span></div><div style="text-align: left;"><span style="font-size: large;"><b><span style="color: #2b00fe;"><br /></span><span style="color: #2b00fe;">Achieving this goal necessitates utilizing <a href="https://blog.modernmt.com/understanding-adaptive-machine-translation/">an adaptable machine translation technology</a> that consistently learns and enhances itself, allowing corporations to interact with individuals in developing economies worldwide.</span></b></span></div><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto;"><tbody><tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhNVQEVXIQ5xQ4w2qANxEDGltOYrQR7Rq4ujzJuI6J_3OMvL6GSmKwwv9J-Wk6l4q7-xgPZrtiRPks39G8s6PzldH5zyNX2dlfziTTYUDuEtZTfIVmBSH2ptrgprGBEi-7q4nUG48WdkZC0Q9QB3kV5buVXFBMPB1gV52wsYupRWM-hFiJ-LeSVzaYeTREM/s2362/Risorsa-22.png" imageanchor="1" style="margin-left: auto; margin-right: auto;"><img border="0" data-original-height="1296" data-original-width="2362" height="220" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhNVQEVXIQ5xQ4w2qANxEDGltOYrQR7Rq4ujzJuI6J_3OMvL6GSmKwwv9J-Wk6l4q7-xgPZrtiRPks39G8s6PzldH5zyNX2dlfziTTYUDuEtZTfIVmBSH2ptrgprGBEi-7q4nUG48WdkZC0Q9QB3kV5buVXFBMPB1gV52wsYupRWM-hFiJ-LeSVzaYeTREM/w400-h220/Risorsa-22.png" width="400" /></a></td></tr><tr><td class="tr-caption" style="text-align: center;">Source: The World in 2050 Study by PwC<br /><div style="text-align: left;"><br /></div></td></tr></tbody></table><br /><div><h1 style="text-align: left;"><strong>The Changing Customer Requirements</strong></h1><p>The
primary motivation for translating enterprise content is to enable and
drive international revenue. Translation of all relevant content is
necessary for building an international business. While the translation
tasks related to the product packaging and basic instructions for the
use of a product are mandated, there is now a growing need for
translating more dynamic content that is a combination of both
unstructured internal corporate content and external non-corporate
content such as user impressions, reviews, and feedback from social
media and influencers on the customer experience related to the product
offerings of a company.</p><p>The digital landscape and audience of the 21st century present a challenging environment for modern global enterprises. </p><div style="text-align: left;"><span style="color: #2b00fe; font-size: large;"><b>Most
internet users do not search for products, but rather they seek answers
to questions. If an enterprise can provide useful answers, potential
customers may develop trust and possibly become advocates and product
champions.</b></span></div><p></p><p>If the enterprise can help a customer
understand, and educate them on the general subject domain, not just
specific product-related subject matter, potential customers may begin
to trust your corporate content, and if you can provide a good customer
experience after they buy your product, they may even advocate using
your products.</p><p><strong>Generally, no potential customer is
searching and hoping to find a website that is merely a corporate ad,
filled with self-congratulating content on how great the company thinks
it is and how wonderful it thinks its products are</strong>.</p><p>Unfortunately, this kind of crap content is still common on many corporate websites which are filled with marketing-speak.<em> S</em>ocial
media technologies have facilitated an ongoing, real-time dialogue that
has reversed the traditional direction of conversations between brands
and their customers. </p><div style="text-align: left;"><span style="color: #2b00fe; font-size: large;"><b>Consumers are now leading the conversation and brands need to listen. </b></span></div><p></p><div style="text-align: left;"><span style="color: #2b00fe; font-size: large;"><b>Trusted,
authentic user-generated content (UGC) has a significant influence on
purchasing decisions, especially in online marketplaces. </b></span></div><p></p><p>New buyers trust the shared authentic experience of other real customers more than any slick pitch created by the corporation. </p><div style="text-align: left;"><span style="color: #2b00fe; font-size: large;"><b>Customers want to know what does not work well, as much as what does, BEFORE they buy.</b></span></div><p></p><p><strong>Marketing-speak
refers to the use of clichés, buzzwords, and empty superlatives in
marketing content. It is typically self-congratulating in tone and is
often an abundance of empty assertions of being “the best” without
meaningful context and trustworthy references. </strong></p><p>Marketing-speak or corporate-speak is often found in press releases, brochures, white papers, and sales letters.</p><p><br /></p><h1 style="text-align: left;">Changing Macroeconomic Trends</h1></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg_ikkCgoUfEmrYMWi3Ln9CllqfuCvXKRqtH6C1jzSwEUnuVBCKDjnOuKTmIhLVIIfh63zI4Vcs6oRXViEU5LSYjqJj6-HX23w-01HrQ7s5btO1U6LsNBH2yC5o5uuH1nLnLfnWQqptUc_WXkiHMnb2P0mSU4KSS7NAIdUNL1rGXc2K2898Ro4r0ETztNDJ/s960/GS%20View.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="225" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEg_ikkCgoUfEmrYMWi3Ln9CllqfuCvXKRqtH6C1jzSwEUnuVBCKDjnOuKTmIhLVIIfh63zI4Vcs6oRXViEU5LSYjqJj6-HX23w-01HrQ7s5btO1U6LsNBH2yC5o5uuH1nLnLfnWQqptUc_WXkiHMnb2P0mSU4KSS7NAIdUNL1rGXc2K2898Ro4r0ETztNDJ/w400-h225/GS%20View.png" width="400" /></a></div><p>The world is also changing both in terms of geographical shifts in
international trade opportunities and relative economic power. The
relative importance of emerging and developing economies is growing, and
any global enterprise wishing to be relevant 10 years from now needs to
understand this shift. </p><p style="text-align: left;"><span style="font-size: large;"><b><span style="color: #2b00fe;">The global economy is gradually moving away from a G7-dominated perspective. Examining these patterns clarifies why this language expansion project is crucial at this moment.</span> </b></span></p><p></p><p>This comparison between the G7 and E7 economies done by <a href="https://www.pwc.com/gx/en/research-insights/economy/the-world-in-2050.html?ref=blog.modernmt.com">Pricewaterhouse Coopers</a> (PwC) for “The World in 2050” study, provides a capsule view of these trends. <strong>PwC estimates six of the seven largest economies in the world are projected to be emerging economies in 2050 </strong>led by China (1st), India (2nd), and Indonesia (4th).</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiZoH3WrAmOx4IFeySzwnU9xcE8lugyeoOdJG5zgm0KJsPqcbefogjCPyXkLhGFStuPgF0Q00_assjXHsZLRnUxGn4yaPlNxFG0t1SlC0XiDKDwQod19oKAZsm43JQRwZ_x2UT9uI9vJNXEqU0W6I0ewwzZGKE38cClr_hUz1yVsENSXdC9j7XRa4dwwUJD/s2347/Risorsa-27.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1237" data-original-width="2347" height="211" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiZoH3WrAmOx4IFeySzwnU9xcE8lugyeoOdJG5zgm0KJsPqcbefogjCPyXkLhGFStuPgF0Q00_assjXHsZLRnUxGn4yaPlNxFG0t1SlC0XiDKDwQod19oKAZsm43JQRwZ_x2UT9uI9vJNXEqU0W6I0ewwzZGKE38cClr_hUz1yVsENSXdC9j7XRa4dwwUJD/w400-h211/Risorsa-27.png" width="400" /></a></div><br /><div>Over the coming decades, emerging economies will drive global growth.
Vietnam, India, and Bangladesh could be three of the fastest-growing
larger economies over this period. <strong>This growth momentum directly
relates to the languages that are growing in importance. By 2050, PwC
projects that the G7’s share of world GDP will fall to only around 20%,
while the E7 will increase their share to almost 50% of global GDP at
Purchasing Power Parity (PPPs). </strong>This means that the top 10 Indic languages which include Bengali become strategic growth opportunities.</div><div><br /></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiTv76lg8JZG-cQr54A0Vngo0MeYY_p3wLd3UxBN7v2ysNM9TvfESZQLBRO4Y9zAzTwckfCZqQPfxFv82TZxyUs8AulXR211nLB8epq5SPsuHoFb_Z67WlY-SwNLsshYZlXAIY5kw8pOWT_52bKeQyNPoubQzpIOPUJ7HU2_RyHUowmhmM6cqC55QEyaPQC/s894/200-LPs-Overview-2.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="685" data-original-width="894" height="306" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiTv76lg8JZG-cQr54A0Vngo0MeYY_p3wLd3UxBN7v2ysNM9TvfESZQLBRO4Y9zAzTwckfCZqQPfxFv82TZxyUs8AulXR211nLB8epq5SPsuHoFb_Z67WlY-SwNLsshYZlXAIY5kw8pOWT_52bKeQyNPoubQzpIOPUJ7HU2_RyHUowmhmM6cqC55QEyaPQC/w400-h306/200-LPs-Overview-2.jpg" width="400" /></a></div><br /><div>The chart above shows estimates of three of the fastest-growing
economies with expected improvements in global rankings. In contrast,
three of the fastest-falling countries are Australia (from 19<sup>th</sup> to 28<sup>th</sup>), Italy (12<sup>th</sup> to 21<sup>st</sup>), and Spain (15<sup>th</sup> to 26<sup>th</sup>).</div><div><p>Some other highlights from the <a href="https://www.pwc.com/gx/en/world-2050/assets/pwc-world-in-2050-slide-pack-feb-2017.pdf?ref=blog.modernmt.com">PwC “The World in 2050” research</a> include:</p><ul><li><strong>The
top 15 fastest-growing economies over the next 30 years will all be
developing and emerging market economies according to PwC projections</strong></li><li>Europe’s share of the world economy at PPPs could fall from around 15% to 9% by 2050</li><li>Brazil and Mexico could be larger than Japan and Germany by 2050</li><li>India could increase its share of world GDP at PPPs by 8% to 15% by 2050</li><li>China’s share of world GDP at PPPs could increase to around 20% by 2050</li></ul><p><strong>Rising
incomes in emerging markets will open up great opportunities for
businesses with sufficiently flexible and patient strategies for these
fast-evolving markets.</strong> As the purchasing power of an increasing
portion of the population grows in these regions, so will the
consumption as we see a rising middle-class emerge.</p><p>Other research
also points to the rise of Asian economies, especially in South Asia
and South East Asia as Chinese GDP growth also starts to slow down as
the demographic impact of the One Child policy kicks in over the next
two decades.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEghSopx8hwOb1ew3tPfOZUeXEZL5MPqGE-g1GORFCNe_UsUPUSRId94mP5kzi-drrY7N5r7Z13VbtkfCsolLM0WcqiHpmLMgG4xjbb7unmvsvv0iUtGyEPWVCbBCEATA1YxXMpJ4ZHgAaDsrcDai16QuhCTIn53QMkDLNquxw9PvZxR4CBZmfEfi7SnpOeD/s792/200-LPs-Overview2.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="538" data-original-width="792" height="271" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEghSopx8hwOb1ew3tPfOZUeXEZL5MPqGE-g1GORFCNe_UsUPUSRId94mP5kzi-drrY7N5r7Z13VbtkfCsolLM0WcqiHpmLMgG4xjbb7unmvsvv0iUtGyEPWVCbBCEATA1YxXMpJ4ZHgAaDsrcDai16QuhCTIn53QMkDLNquxw9PvZxR4CBZmfEfi7SnpOeD/w400-h271/200-LPs-Overview2.jpg" width="400" /></a></div><br /><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjZZ5uFtV5QrrvILERSNfpjHWSzBKpumgRIWRd6xNy5AnDGOK3Mw6P6JlU_yzgeQ5AQ3JfEnu_FM0w5HgVnKSGzpkxAwQpBpgWqTFsyLswzo1lqvUi2RBS_k8O6nvvRWX03uQa55fInSw6n3Ubo4pfuabCe1UJbYuAA4XPCIV_1sZUGs3Rx57cdrtj21Ou9/s805/200-LPs-Overview-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="649" data-original-width="805" height="323" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjZZ5uFtV5QrrvILERSNfpjHWSzBKpumgRIWRd6xNy5AnDGOK3Mw6P6JlU_yzgeQ5AQ3JfEnu_FM0w5HgVnKSGzpkxAwQpBpgWqTFsyLswzo1lqvUi2RBS_k8O6nvvRWX03uQa55fInSw6n3Ubo4pfuabCe1UJbYuAA4XPCIV_1sZUGs3Rx57cdrtj21Ou9/w400-h323/200-LPs-Overview-1.jpg" width="400" /></a></div><p>Goldman Sachs was one of the first to point out the emerging rise of the
BRIC economies 15 years ago. This forecast has been more accurate for
the Asian economies but in their <a href="https://www.goldmansachs.com/insights/pages/gs-research/the-path-to-2075-slower-global-growth-but-convergence-remains-intact/report.pdf?ref=blog.modernmt.com">latest research</a>, they expect that growth will be more evenly distributed even though Asian economies will still dominate.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEheFibmxy-0vLmmkjj70rsFmL1E832qgpIWxh2POehloTXaYxNw7ygovnjPuVYN-i-VY-yoGA5hPqgg681kKq9DMpN6AM1UMkXIDjMcSC2OAH3cY7OubPOPMAeYongM2R1BZ_ceAayhYvnVFSfOKI-898uyUvbIR6Rlpjk5ElG87bROL57XNVO_vm8rNe2W/s1119/200-LPs-Overview4.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="589" data-original-width="1119" height="210" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEheFibmxy-0vLmmkjj70rsFmL1E832qgpIWxh2POehloTXaYxNw7ygovnjPuVYN-i-VY-yoGA5hPqgg681kKq9DMpN6AM1UMkXIDjMcSC2OAH3cY7OubPOPMAeYongM2R1BZ_ceAayhYvnVFSfOKI-898uyUvbIR6Rlpjk5ElG87bROL57XNVO_vm8rNe2W/w400-h210/200-LPs-Overview4.jpg" width="400" /></a></div><p><strong>China, Vietnam, Uganda, Indonesia, and India are projected to be among the fastest-growing economies by 2030 according to the <a href="https://atlas.cid.harvard.edu/growth-projections?ref=blog.modernmt.com">Harvard Growth Lab projections</a>.</strong>
Their research also factors in the ability of the country to develop
complex production capabilities and finds that countries that have
diversified their production into more complex sectors, like Vietnam and
China, are those who will experience the fastest growth in the coming
decade.</p><p>The Harvard Growth Lab has identified three growth poles using their <a href="https://atlas.cid.harvard.edu/growth-projections?ref=blog.modernmt.com">Economic Complexity Index</a> (ECI) which they believe is a much better predictor of economic growth prospects. </p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgViz1s7Adr73LAje1sFjV3M1rfRwxe1d_4APbC-N7--wHmfIG_WAIqQYYdoAmPQw6bRjTfG9abz1YEow_OIJNj0dIPdWqJPJV7hXbMjfces4fatxoHVzq68renuYfSD9_mgqwJvXfcGVnpd1WvKvB078OX0A8e6LiiYr251LRyCncEp7V7n_uh19G__dQc/s1079/200-LPs-Overview2-1.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="837" data-original-width="1079" height="310" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgViz1s7Adr73LAje1sFjV3M1rfRwxe1d_4APbC-N7--wHmfIG_WAIqQYYdoAmPQw6bRjTfG9abz1YEow_OIJNj0dIPdWqJPJV7hXbMjfces4fatxoHVzq68renuYfSD9_mgqwJvXfcGVnpd1WvKvB078OX0A8e6LiiYr251LRyCncEp7V7n_uh19G__dQc/w400-h310/200-LPs-Overview2-1.jpg" width="400" /></a></div><p>They state that several Asian economies already hold the necessary
economic complexity needed to drive the fastest growth over the coming
decade, led by <strong>China, Cambodia, Vietnam, Indonesia, Malaysia, </strong>and <strong>India</strong>.
In East Africa, several economies are expected to experience rapid
growth, though this is driven more by population growth than gains in
economic complexity, and this includes <strong>Uganda, Tanzania, </strong>and <strong>Mozambique</strong>.
They also saw several Eastern European countries including Georgia,
Lithuania, Belarus, Armenia, Latvia, and Romania ranking high on a per
capita basis because of improvements in their ECI.</p><p>According to the IMF’s recent World Economic Outlook on Africa, five of the world’s fastest-growing economies are <strong>Angola, Ethiopia, Nigeria, Kenya, and South Africa. </strong>However,
many experts say that for most of Africa, the business opportunity for
global enterprises is further out into the future. Perhaps in the
10-to-20-year time frame as a properly supported (improved
infrastructure, education, health services) demographic dividend starts
to kick in. However, there are some exceptions as shown in the Growth
Lab chart above.</p><p>This growth momentum data correlates with
macroeconomic business potential, but a global business needs to
determine a market opportunity match by including several other factors
beyond the possibility that there is a large and growing potential
customer base. Apart from product fit and basic organizational
infrastructure issues needed to serve global customers across the world,
several other macroeconomic factors also need to be considered. These
include all of the following:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEizmhPPP5gyuSwvpTRA2UDyqNgJbpyOM5NaDlHjtFI2Sfjcgeb5TVYpFNuHGe8SQeYSS1S-n9WMLipB6lZHDtBdp7wT_fjvhTf6nUuWZx5jzsvneWNk5z6LQLqy13wil0n9wY4nSV2CWEg3dJY50YaGV0pl9jFUTO8hIgGtIfXoQxU-FNGf4JLGTFE1Wx0x/s1131/ModernMT-Expands-Language-Coverage-to-200-Languages-04.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="479" data-original-width="1131" height="170" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEizmhPPP5gyuSwvpTRA2UDyqNgJbpyOM5NaDlHjtFI2Sfjcgeb5TVYpFNuHGe8SQeYSS1S-n9WMLipB6lZHDtBdp7wT_fjvhTf6nUuWZx5jzsvneWNk5z6LQLqy13wil0n9wY4nSV2CWEg3dJY50YaGV0pl9jFUTO8hIgGtIfXoQxU-FNGf4JLGTFE1Wx0x/w400-h170/ModernMT-Expands-Language-Coverage-to-200-Languages-04.png" width="400" /></a></div><br /><p>After this analysis has been done the best-fitting products and services can be presented to the new market. <strong>Market
viability tests can often be done initially by creating a digital
presence and window front to assess interest and better define
implementation issues.</strong><strong> </strong></p><div class="separator" style="clear: both; text-align: center;"><strong><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjIyZp_NKMAUnqYDJoYRkgKIVKhraPP9RzjdQXzFYFs443p46dXUXPB2z_WXyWdOzUUfQBvOyB_Lh1ut7NZQieyydFIcA2ynkp0vOS7H86XUpCdxEM_Y1jREg3ZKgV_kEXJm34gIOgHrU2td1CuIB0ZoF7xnrnmqclB3a-fDi26oZI6R47b0H-uS997_QLS/s1020/ModernMT-Expands-Language-Coverage-to-200-Languages-05.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="251" data-original-width="1020" height="99" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjIyZp_NKMAUnqYDJoYRkgKIVKhraPP9RzjdQXzFYFs443p46dXUXPB2z_WXyWdOzUUfQBvOyB_Lh1ut7NZQieyydFIcA2ynkp0vOS7H86XUpCdxEM_Y1jREg3ZKgV_kEXJm34gIOgHrU2td1CuIB0ZoF7xnrnmqclB3a-fDi26oZI6R47b0H-uS997_QLS/w400-h99/ModernMT-Expands-Language-Coverage-to-200-Languages-05.png" width="400" /></a></strong></div><strong><br /></strong><p></p><p><strong>It is at this point that the relevant content for the buyer
and customer journeys and translation issues come into focus. This is
what ModernMT and the service and process infrastructure at Translated
are designed to address.</strong></p><h1 style="text-align: left;"><strong>The ModernMT Language Expansion</strong></h1><p>One
of the motivating ideas behind Translated's language expansion efforts
is to help our enterprise customers reach more of the world's
population. By expanding language coverage to potentially reach 6.5
billion native speakers, Translated enables companies to forge stronger
connections with global users and customers by building the core
translation infrastructure needed to share, communicate, and listen to
new customer groups.</p><p><strong>This initial launch and introduction of these new languages is the beginning of an evolutionary process</strong>.
Translated has ensured that the current quality of MT produced by these
new languages is at least equal to or better than systems produced by
Big Tech companies, and has added high-quality training data resources
when available to immediately improve the performance of “low-resource”
languages.</p><p><strong>The expectation and plan behind this launch are to enable these language systems <a href="https://blog.modernmt.com/understanding-adaptive-machine-translation/">to start improving immediately using the highly adaptive ModernMT technology</a> which allows this to happen.</strong> </p><p>History
shows that because of the volume of production work, greater data
availability, and ongoing activity around the high-resource languages,
those MT systems can reach levels of accuracy where discussions of
human-equivalent performance are possible. Thus, we begin this journey
with a large set of new languages. In addition to this gradual
improvement effort, ongoing fundamental research will continue to
increase the rate at which these systems can and will improve.</p><p style="text-align: left;"><span style="color: #2b00fe; font-size: large;"><b>Digital
leadership in emerging markets will require that enterprises translate
tens of millions of words a month, to enable them to listen,
communicate, understand, and share relevant information with these
customers.</b></span></p><p></p><p>For the first time, 30 new languages are
supported in the market, leapfrogging directly to the more powerful
adaptive MT technology. Among the new languages now supported by
ModernMT, are Bengali, Punjabi, and Javanese, which together with all
the other newly added languages are spoken by over 2 billion people
worldwide. Many of these languages have high commercial potential,
enabling companies to connect and engage with some of the
fastest-growing economic regions in the world.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhzs_KmvwDd_uVic8PRIGQ3UojiA5I31cECs6JxXRSOBjmxNcls200s8sXsMrdeHZt7mEbkGp7d9yGrc3B8NXxPSbt4EnQ13Wuh32ka825yhfvbEHioMhiR4tPU59GxjxFAYJShpNOyyit1fqKO6qbg7314GXxXOlMLnIYdNs8rs0DVBbdu0nHWvtyIDKhu/s1017/ModernMT-Expands-Language-Coverage-to-200-Languages-06.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="306" data-original-width="1017" height="120" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhzs_KmvwDd_uVic8PRIGQ3UojiA5I31cECs6JxXRSOBjmxNcls200s8sXsMrdeHZt7mEbkGp7d9yGrc3B8NXxPSbt4EnQ13Wuh32ka825yhfvbEHioMhiR4tPU59GxjxFAYJShpNOyyit1fqKO6qbg7314GXxXOlMLnIYdNs8rs0DVBbdu0nHWvtyIDKhu/w400-h120/ModernMT-Expands-Language-Coverage-to-200-Languages-06.png" width="400" /></a></div><br /><p>In the customer-centric world of the future, it is also important
that important tools in the localization technology stack can easily
interact and connect to <a href="https://blog.modernmt.com/understanding-adaptive-machine-translation/">superior continuous learning tools like ModernMT.</a> To further enable this we have also added API connectivity to <a href="https://www.blackbird.io/?ref=blog.modernmt.com">Blackbird.io</a>
which is an Integration Platform as a Service (IPaaS). The
inter-application connectivity reach will continue to improve. This will
allow ModernMT to ingest data from and export translated data back to a
growing set of TMS, CMS, Marketing Automation, CDP, Analytics, QE, and
Storage solutions needed in modern CX-related automation deployments. </p><p><strong>Unfortunately, many TMS systems of yesteryear still have no ability to interact with <a href="https://blog.modernmt.com/understanding-adaptive-machine-translation/">fast-evolving adaptive MT systems</a> and trap enterprise data to create tech debt that undermines global success. Buyers need to be wary of such systems and move to open-source approaches that allow the greatest flexibility and agility.</strong></p><p>Marco Tombetti was <a href="https://multilingual.com/modernmt-marco-qa/?ref=blog.modernmt.com">interviewed by Multilingual </a>about
this announcement, where he explains how data scarcity, and enabling
the new languages to function with adaptive MT architecture were the two
main challenges that had to be overcome to achieve this.</p><p>He also said, <strong><em>“We envision that this effort is merely the
first step, and while 200 languages may appear substantial, it is not an
extraordinary figure. We are at the beginning, and <a href="https://jlmr-zgpvh.maillist-manage.net/click/19d1d8cf00784053/19d1d8cf00783fc0?ref=blog.modernmt.com">we plan to refine adaptive MT support</a> for these languages in the coming months, as well as for numerous others.</em></strong>”</p><h3 id="for-a-full-listing-of-the-languages-supported-by-modernmt-httpswwwmodernmtcomapilanguages">For a full listing of the languages supported by ModernMT: <a href="https://www.modernmt.com/api/?ref=blog.modernmt.com#languages">https://www.modernmt.com/api/#languages</a></h3><p><br /></p></div>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0tag:blogger.com,1999:blog-6748877443699290050.post-28921842648032841602023-05-30T13:52:00.003-07:002023-05-30T14:13:26.346-07:00Understanding Adaptive Machine Translation<p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Machine translation has been around for over 70 years and has made steady progress in tackling what many consider to be one of the most difficult challenges in computing and artificial intelligence. We have seen the approach to this challenge change and evolve, and MT has become much more widely used, especially since the advent of neural MT.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">The deep learning neural net methods used in Neural MT have led to significant improvements in output quality, especially in terms of improved fluency, and have encouraged much wider use of machine translation in global business-driving applications.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">While the momentum of Neural MT is well understood and recognized as a major advance in state-of-the-art (SOTA) machine translation, it is surprising that Adaptive MT has not had a greater impact.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">This is especially true in the enterprise and professional translation market, where Adaptive MT can address specific and unique business needs much more effectively than alternatives. This paper explains why, even in the age of large language models, it remains a critical technology for global enterprises and professional users.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="color: #2b00fe; font-size: medium;">To better understand the value of Adaptive MT systems, it is useful to present a contrast to the typical generic static systems that most are familiar with.</span></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"></span></p><h3 style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 20px; font-weight: 500; line-height: 1.1; margin-bottom: 10px; margin-top: 20px; text-align: center;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">The Typical Generic Static MT Engine</span></h3><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgEcbN_zDLu-VMQvyuHSi4LHsmFzCf52p7v1lR-xia7_iUJmRWxApDn1h7A2QrFX6F5_ADaKDA4WJapYePtlQ_M-njurVqcFcqhizqGX_hQWu-5AydOXxEpGL9htn6qtB9fbfOv612j9VC2sWeyaycmiedvaX5fXnxKVdJlJ-8aDWGb1jx0MKx43xemWg/s2958/Understanding-Adaptive-Machine-Translation_1%20(1).png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1125" data-original-width="2958" height="153" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgEcbN_zDLu-VMQvyuHSi4LHsmFzCf52p7v1lR-xia7_iUJmRWxApDn1h7A2QrFX6F5_ADaKDA4WJapYePtlQ_M-njurVqcFcqhizqGX_hQWu-5AydOXxEpGL9htn6qtB9fbfOv612j9VC2sWeyaycmiedvaX5fXnxKVdJlJ-8aDWGb1jx0MKx43xemWg/w400-h153/Understanding-Adaptive-Machine-Translation_1%20(1).png" width="400" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><div class="separator" style="clear: both; text-align: left;"><span style="color: #333333; font-family: Helvetica Neue, Helvetica, Arial, sans-serif;"><span style="font-size: 14px;">The MT engines that are generic and relatively static are intended for use by a large number of people. These engines do not alter how a phrase is translated until the next major update. Creating a generic baseline engine requires a considerable amount of effort in terms of cost, complexity, and data, which is why it is not done frequently. The diagram above displays the usual process involved in developing and producing a static engine.</span></span></div><div class="separator" style="clear: both; text-align: left;"><span style="color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px;"><br /></span></div><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="color: #2b00fe; font-size: medium;">A key characteristic of these static engines is that they do not evolve quickly because they require large new data sources to drive improvements, data that is not readily available, and thus generic static engines are typically updated no more than once a year.</span></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">On any given day, the major generic MT engine portals (Google, Microsoft, Baidu) allow hundreds of millions of people to translate material of interest. We have already reached the point where 99% or more of the translation done on the planet is done by computers, thanks to these generic engines.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">In professional or business settings, the demands for using Machine Translation (MT) are quite particular. <b>Generally, generic MT engines need to be tweaked and fine-tuned to cater to company- or project-specific language usage and terminology. </b>This process of adjusting the MT engine to suit corporate requirements is known as customization or adaptation.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">For example, if we consider the needs of IKEA, Pfizer, Airbnb, and Samsung, it is clear that they all have very different needs in terms of subject domain focus, style, and critical terminology and would be better served by enterprise-optimized MT than by a generic, one-size-fits-all MT solution.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; background-color: white; box-sizing: border-box;"><span style="color: #2b00fe; font-size: large;">Customization or adaptation of MT models to the correct terminology is necessary for successful outcomes with MT use in most enterprise or professional use settings.</span></span></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">The typical MT customization process using static engines is described below. The customization effort and process is a scaled-down version of the generic engine development process. Typically, it requires the collection and incorporation of enterprise translation memory relevant to the use case into the generic model via a scaled-down "training process."</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">This effort results in limited or coarse optimization if sufficient training data resources are available. The optimization is considered coarse because the training data available to perform the optimization is typically minuscule compared to the base data used in the generic engine. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">There is little value in training an engine with limited data as there would be no difference in performance from the generic baseline.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; background-color: white; box-sizing: border-box;"></span></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Thus, many attempts to use MT in professional settings face data scarcity problems. Limited data availability limits and reduces the potential impact of adaptation.</span> To further complicate matters, it is usually necessary to build separate engines for each different use case, e.g., customer support, marketing, and legal would all be optimized separately.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhNTP54_fVtTjyOV7tVcZzs448Kvz9mkoGeQxL7WCZKnotLTJeXDB7LHPtZhCmMo-K7Zx7sI7gt9u7XIP9sIA78ZQBBIXwjJ537uz-xSGCfFIAKQuj5kcjQizPSAusyd5l8Hz6Z9YnD__7umRe0yi3B1dwEeyMql8tbp35WBzal9yVNQf-XrVDfkKUP8w/s3283/Understanding-Adaptive-Machine-Translation_2-2.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1224" data-original-width="3283" height="149" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhNTP54_fVtTjyOV7tVcZzs448Kvz9mkoGeQxL7WCZKnotLTJeXDB7LHPtZhCmMo-K7Zx7sI7gt9u7XIP9sIA78ZQBBIXwjJ537uz-xSGCfFIAKQuj5kcjQizPSAusyd5l8Hz6Z9YnD__7umRe0yi3B1dwEeyMql8tbp35WBzal9yVNQf-XrVDfkKUP8w/w400-h149/Understanding-Adaptive-Machine-Translation_2-2.png" width="400" /></a></div><div><span face=""Helvetica Neue", Helvetica, Arial, sans-serif" style="background-color: white; color: #333333; font-size: 14px;">Since many global enterprises have multiple product lines and businesses that cross multiple domains (TVs, semiconductors, PCs, home appliances) this will often result in a large number of MT engines needed to cover global business needs. As a result, </span><span face=""Helvetica Neue", Helvetica, Arial, sans-serif" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; background-color: white; box-sizing: border-box; color: #333333; font-size: 14px; font-weight: 700;">it is often necessary to manage and maintain many MT engines. </span><span face=""Helvetica Neue", Helvetica, Arial, sans-serif" style="background-color: white; color: #333333; font-size: 14px;">This management burden is often not understood at the outset when localization teams embark on their MT journey. This complexity also creates a lot of room for error and misalignment as data alignment can easily get out of sync over time.</span></div><div><span face="Helvetica Neue, Helvetica, Arial, sans-serif" style="color: #333333;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhXe54SD747PM-JzA-3KXBh1TCTFTHBv-IFMHERd7hzHqTQDMFjbi2NHB_EGxOTKegWRY0WIdzKpeeya05XAdy_rrX1ckny3aTDa9Euk1Gao64lvU8OcJktsrsR7kSHQCvMS0Fc_BzepPW03W5nP9wcgjZFo5LVUuQ-6ItVI1-A4zwcg3rAZDADFyzTOw/s2729/Understanding-Adaptive-Machine-Translation_3.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1482" data-original-width="2729" height="217" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhXe54SD747PM-JzA-3KXBh1TCTFTHBv-IFMHERd7hzHqTQDMFjbi2NHB_EGxOTKegWRY0WIdzKpeeya05XAdy_rrX1ckny3aTDa9Euk1Gao64lvU8OcJktsrsR7kSHQCvMS0Fc_BzepPW03W5nP9wcgjZFo5LVUuQ-6ItVI1-A4zwcg3rAZDADFyzTOw/w400-h217/Understanding-Adaptive-Machine-Translation_3.png" width="400" /></a></div></span><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Over time, many enterprise MT initiatives can be characterized by several problems that are common to users of these static MT systems. These problems are summarized below in order of frequency and importance:</p><ol style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin-bottom: 10px; margin-top: 0px;"><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Ongoing scarcity of training data: </span>Static models require a lot of data to drive improvements.<span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"> There is little value in retraining a model until new or corrective data volumes reach critical levels.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Tedious MTPE experience: </span>Post-editors must repeatedly correct the same errors because these MT engines do not regularly improve, often leading to worker dissatisfaction.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">MT model management overhead and complexity: </span>There are too many models to manage and maintain, which can lead to misalignment errors.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Communication issues: </span>Typically, between the MT development team and localization team members and translators, who have very different views of the overall process.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Context insensitivity: </span>Sentence- and document-level context is typically missing from these custom models.</li></ol><div><span face="Helvetica Neue, Helvetica, Arial, sans-serif" style="color: #333333;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhatQzWhe2t3fnwlhAkVZ5CZfCsJKd8wzeu1iIVvbbs_SE5U5h0-iO1AuGXdgBTSuwibAW3jNm8rabvSSwlX0I-nSdY5rufhdkw1RFfV9g84lBEzlGNbp-vS2Kta4e8d8EzfIo7heYngKwP5vjfcZl6jdFaCYG51f-bu4TvfQV2l8EIqIM8YECB98kQvA/s2985/Understanding-Adaptive-Machine-Translation_4-1.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1380" data-original-width="2985" height="185" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhatQzWhe2t3fnwlhAkVZ5CZfCsJKd8wzeu1iIVvbbs_SE5U5h0-iO1AuGXdgBTSuwibAW3jNm8rabvSSwlX0I-nSdY5rufhdkw1RFfV9g84lBEzlGNbp-vS2Kta4e8d8EzfIo7heYngKwP5vjfcZl6jdFaCYG51f-bu4TvfQV2l8EIqIM8YECB98kQvA/w400-h185/Understanding-Adaptive-Machine-Translation_4-1.png" width="400" /></a></div><br /></span><span style="color: #2b00fe; font-family: Helvetica Neue, Helvetica, Arial, sans-serif; font-size: medium;"><span style="background-color: white;"><i>From a technical standpoint, static MT systems often have a significant disparity between the encoding (training) and decoding (inference) stages of model deployment, resulting in a notable disconnect. <b>Adaptive MT, on the other hand, aims to bridge the gap between the two phases of model creation (training) and deployment (inference), thus providing more effective support to expert users such as translators and linguists.</b></i></span></span></div><div><em style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; background-color: white; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><br /></span></em></div><h2 style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 25px; line-height: 1.1; margin-bottom: 10px; margin-top: 20px;">The Adaptive MT Experience</h2><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">The <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">static MT approach makes sense for large ad-supported portals</span> where the majority (99%+) of the millions of users will use the MT systems without attempting modification or customization.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">In contrast,<span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"> the adaptive MT approach makes more sense for those enterprise and professional translators who almost always attempt to modify the behavior of the generic model to meet the specific and unique needs of a business use case.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="color: #333333;">ModernMT is an adaptive MT technology solution designed from the ground up to enable and encourage immediate and continuous adaptation to changing business needs.</span><span style="color: #2b00fe;"> <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">It is designed to support and enhance the professional translator's work process and increase translation leverage and productivity. </span></span><span style="color: #333333;">This is the fundamental difference between an adaptive MT solution like ModernMT and static generic MT systems.</span></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVjEMbCK827vCJkRvkhlfomgkN2R_IPtHP_LaMdIcln2rYVxAU6Afasng-5AboBT9aOP444TtLiBlt7Cn6i6KI_YxMh1nYbHUoJlkuNq0v1oj8wZESMINnH9ghynE3SBjkzdSTmXwOKsYY-gnb28vuLwe7wy-HuGUBsaXdIATSGNQlQXxmZQL3_aKYeQ/s3414/Preferred-Updated-Translation_5.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1469" data-original-width="3414" height="173" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiVjEMbCK827vCJkRvkhlfomgkN2R_IPtHP_LaMdIcln2rYVxAU6Afasng-5AboBT9aOP444TtLiBlt7Cn6i6KI_YxMh1nYbHUoJlkuNq0v1oj8wZESMINnH9ghynE3SBjkzdSTmXwOKsYY-gnb28vuLwe7wy-HuGUBsaXdIATSGNQlQXxmZQL3_aKYeQ/w400-h173/Preferred-Updated-Translation_5.png" width="400" /></a></div><br /><h2 data-mce-style="text-align: center;" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 25px; font-weight: 500; line-height: 1.1; margin-bottom: 10px; margin-top: 20px; text-align: center;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="color: #2b00fe;">“Simplicity is the ultimate sophistication”</span></span></h2><p data-mce-style="text-align: center;" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px; text-align: center;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="color: #2b00fe;">Leonardo da Vinci</span></span></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiY5EHlnxbuNfxLbuUkWROkNwCX9CxERxftog3kRDOGFTwxFFvhhsG6cgUa-uHO5IsurmFZQDCa-ZoeeRi_xX1koOtaRGdy_UwGvZTAmInPA3-WH47nMB7nJ0XBDue2eTYv0QfWuguNVc94XGPHrnEBIjQgAvjqe99hnKCpSgJZL8mVWNoNJCDN_IxI8Q/s2662/Understanding-Adaptive-Machine-Translation_6.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1479" data-original-width="2662" height="223" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiY5EHlnxbuNfxLbuUkWROkNwCX9CxERxftog3kRDOGFTwxFFvhhsG6cgUa-uHO5IsurmFZQDCa-ZoeeRi_xX1koOtaRGdy_UwGvZTAmInPA3-WH47nMB7nJ0XBDue2eTYv0QfWuguNVc94XGPHrnEBIjQgAvjqe99hnKCpSgJZL8mVWNoNJCDN_IxI8Q/w400-h223/Understanding-Adaptive-Machine-Translation_6.png" width="400" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">While the ModernMT adaptive MT engine also has a basic generic engine underlying its capabilities, <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">it is designed to work instantly with any available translation memory resources and to learn instantly from corrective linguistic feedback.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">This is done without any user intervention or action to "train" the system. The user simply points to any available TM and <b>it is used if it is relevant to the translation task at hand.</b> Thus, while many struggle to use MT in an environment where use case requirements are constantly changing, <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">this adaptive MT system uses memories, corrective feedback, and overall context gathered from both the memories and the overall document.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgmTpc-YebJqGXhLAAJOUE8owftnWGUUIE-eAcMkR8ko7MHWpmUZJkgLwh3IFVi0WttIT3MPtW6qv781St6Yf_5a4auRn9qUzYkhxqcr61qhlwv16Ho73bNJBGnr2oAof88AXs4emvuoHoHWp2gaEtTu1E-oIFxCII67cPQXwq1xeq1oCuPpw1-oHkfqQ/s3355/Understanding-Adaptive-Machine-Translation_7.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1520" data-original-width="3355" height="181" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgmTpc-YebJqGXhLAAJOUE8owftnWGUUIE-eAcMkR8ko7MHWpmUZJkgLwh3IFVi0WttIT3MPtW6qv781St6Yf_5a4auRn9qUzYkhxqcr61qhlwv16Ho73bNJBGnr2oAof88AXs4emvuoHoHWp2gaEtTu1E-oIFxCII67cPQXwq1xeq1oCuPpw1-oHkfqQ/w400-h181/Understanding-Adaptive-Machine-Translation_7.png" width="400" /></a></div><br /><p></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">As the use of MT grows in the enterprise, the benefits of an adaptive MT infrastructure continue to accrue, as the management and maintenance of the many production MT systems require nothing more than the organization of TM assets and the provision of continuous corrective feedback to drive continuous improvements in system performance.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Thus, content creators and linguistically informed users can be the primary drivers of the ongoing system evolution. Because the underlying continuous improvement process is always active in the background, there is no need for any technology process management by these users.<span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"> Translation issues that may arise in widespread use, can be quickly identified and corrected by linguists without the need for support from MT technology experts.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">New use cases of large-scale deployments can be rapidly deployed by targeting human translation efforts on the most relevant and statistically present content. Adaptive MT technology allows for evolutionary approaches that ensure continuous improvement.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjBGtwyf3uyCiu1_kmUCWDqMCDJ3vmFvMdwWSHHuW-gmABHqnlygPrTprUbR5UrZyFNo_fOffBf94dR4r98UPq4de3LUvuXcSG2UR5x3DYjiZ4-GjJBx9z7D7PkvO4rMxRKz3uNJLlt39B626NhqhdhgUejbdtMHLZRnz8NjwLWi45GFkOaJMqqm4s4Bw/s3025/Understanding-Adaptive-Machine-Translation_8.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1570" data-original-width="3025" height="208" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjBGtwyf3uyCiu1_kmUCWDqMCDJ3vmFvMdwWSHHuW-gmABHqnlygPrTprUbR5UrZyFNo_fOffBf94dR4r98UPq4de3LUvuXcSG2UR5x3DYjiZ4-GjJBx9z7D7PkvO4rMxRKz3uNJLlt39B626NhqhdhgUejbdtMHLZRnz8NjwLWi45GFkOaJMqqm4s4Bw/w400-h208/Understanding-Adaptive-Machine-Translation_8.png" width="400" /></a></div><br /><p></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Independent market research points to some key factors that are often overlooked by those attempting to deploy MT in professional and enterprise environments. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Surveys conducted by Common Sense Advisory and Nimdzi show that most LSPs/Enterprises struggle to deploy MT in production </span>for three key reasons:</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"></span></p><ol style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin-bottom: 10px; margin-top: 0px;"><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Inability to produce MT output at the required quality levels</span>. Most often due to a lack of training data needed for meaningful improvement.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Inability to properly estimate the effort and cost of deploying MT</span> in production.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">The ever-changing needs and requirements of different projects </span>with static MT that cannot adapt easily to new requirements create a mismatch of skills, data, and competencies.</li></ol><div><span face="Helvetica Neue, Helvetica, Arial, sans-serif" style="color: #333333;"><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiWefwGT-yTWnC8TPGtSRUME_XaQFnKucGGAiQk9xeU93plAM2UlKEk-VjQMBGQirCamMl5PCVoeY3L2wTTz5SeSQOvt7rUhGBfCMmZpf1RTBH4wTpYvp2sy95CSYTGMZabf6kf_vqP3_9mh_skgTPV8VIb0eNuPrPR7gxp-jhbM1eAIx4UrFiIuyJlUA/s2898/Understanding-Adaptive-Machine-Translation_9.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1551" data-original-width="2898" height="214" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiWefwGT-yTWnC8TPGtSRUME_XaQFnKucGGAiQk9xeU93plAM2UlKEk-VjQMBGQirCamMl5PCVoeY3L2wTTz5SeSQOvt7rUhGBfCMmZpf1RTBH4wTpYvp2sy95CSYTGMZabf6kf_vqP3_9mh_skgTPV8VIb0eNuPrPR7gxp-jhbM1eAIx4UrFiIuyJlUA/w400-h214/Understanding-Adaptive-Machine-Translation_9.png" width="400" /></a></div></span><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Given these difficulties, it is worth considering the key requirements for a production-ready MT system. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Why do so many still fail with MT?</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">One reason for failure is that many LSPs and localization managers have used automated metrics to select the "best" MT system for their production needs without having any understanding of how MT engines improve and evolve. </span>Automated MT quality metrics such as BLEU, Edit Distance, hLepor, and COMET are used to select the "best" MT systems for production work.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">The scores can be helpful for MT system developers in enhancing and refining their systems. However, globalization managers who rely solely on this method to choose the "best" system may miss some noticeable limitations in selecting the most suitable MT system.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">Ideally, the "best" MT system would be determined by a team of competent translators who would run</span> <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">directly relevant content through the MT system after establishing a structured and repeatable evaluation process. </span></span>This is slow, expensive, and difficult, even if only a small sample of 250 sentences is evaluated.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Thus, automated measurements (metrics) that attempt to score translation adequacy, fluency, precision, and recall must often be used. They attempt to do what is best done by competent bilingual humans.<span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"> These scoring methodologies are always an approximation of what a competent human assessment would determine, and can </span>often<span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"> be incorrect or misleading</span>, especially with static Test Sets.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">This approach of ranking different MT systems by scores based on opaque and possibly irrelevant reference test sets has several problems. These problems include:</p><ul style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin-bottom: 10px; margin-top: 0px;"><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">These scores do not represent production performance.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">These scores are <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">typically obtained on static MT systems and do not capture a system's ability to improve.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">The results are an OLD snapshot of a constantly changing scene</span>. If you change the angle or focus, the results would change.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">Small differences in scores are often meaningless</span>, and most users would be hard-pressed to explain what these small numerical differences might mean.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The score is an approximate measure of system performance at a historical point in time and is generally not a reliable predictor of future performance.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">These scores are unable to capture the dynamic evolution typical of an adaptive MT system.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">Generic, static systems often score higher on these rankings initially but this does not reflect that they are much more difficult to tune and adapt to unique, company-specific requirements.</li></ul><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><b>When choosing MT systems for production use, relying solely on score-based rankings can lead to suboptimal or even incorrect choices. </b>This approach is often used because NMT system performance is difficult to understand and can be shrouded in mystery. However, automated metrics are not always reliable and should not be the only factor considered in making purchase decisions. Using scores to justify choices may be a "lazy buyer" strategy that fails to fully account for the complexity involved in selecting the best MT system for a given purpose.</p></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjIrpBtvxEr0tD8OHBhhpHeX9v6qH4-Ql4SEl73ljJZLSJhZlDuo7hssxXS0dV4sTWvFlgPTbV6_IiQ8yVEg6jC1g1h_myaOcJxsR2Zxv16DBJtNh4gYc0KJ1N6GKKM7aXpy7SFS72HyG94zmw_DShh2yHQlj3hELAOtVb6VbYdziw-L0ZUbFbHwJy1Ag/s3492/Understanding-Adaptive-Machine-Translation_10.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1347" data-original-width="3492" height="154" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjIrpBtvxEr0tD8OHBhhpHeX9v6qH4-Ql4SEl73ljJZLSJhZlDuo7hssxXS0dV4sTWvFlgPTbV6_IiQ8yVEg6jC1g1h_myaOcJxsR2Zxv16DBJtNh4gYc0KJ1N6GKKM7aXpy7SFS72HyG94zmw_DShh2yHQlj3hELAOtVb6VbYdziw-L0ZUbFbHwJy1Ag/w400-h154/Understanding-Adaptive-Machine-Translation_10.png" width="400" /></a></div><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">But the failure of so many LSPs with MT technology suggests that this approach may not be the best way forward to achieve production-ready and production-grade MT technology. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">So what criteria are more relevant in the context of identifying production-grade MT technology?</span> The following criteria are much more likely to lead to technology choices that make long-term sense. For example:</p><ul style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin-bottom: 10px; margin-top: 0px;"><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">speed with which an MT system can be tuned and adapted</span> to unique corporate content. Systems that require complex training efforts by technology specialists will slow the globalization team’s responsiveness.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">ease with which the system can be adapted</span> to unique corporate needs The need to have expensive consulting resources or dedicated MT technology staff on hand and ready to go greatly reduces the agility and responsiveness of the globalization team.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">An <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">automated and robust MT model improvement process </span>as corrective feedback and improved data resources are brought to bear.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">complexity of MT system management</span> increases exponentially when multiple vendors are used as they may have different maintenance and optimization procedures. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">This suggests that it is better to focus on one or two partners and build expertise through deep engagement.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The ability of a system to <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">enable startup work even if little or no data is available.</span></li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">A <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">straightforward process to correct any problematic or egregious translation errors</span>. Many large static systems need large volumes of correction data to override such errors.</li><li style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box;">The <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">availability of expert resources to manage specialized enterprise use cases</span> and trained human resources (linguists) to help prime and prepare MT systems for large-scale deployment.</li></ul><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">It is now common knowledge that machine learning-based AI systems are only as good as the data they use. <span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">One of the keys to long-term success with MT is to build a virtuous data collection system that refines MT performance and ensures continuous improvement.</span> The existence of such a system would encourage more widespread adoption and enable the enterprise to become multilingual at scale. This would allow the enterprise to break down the barrier of language as a barrier to global business success.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiOnlF3lWWnYxRD6d4jrbW1X4Z22Bihbt6KJUqt5H36RVH_pfA9OQGjDz4s22M6la1yI0WkpbGLuXoMVK8Q1C0TiD_2mLnp678nDoqwp1D_g-wq76_rhWKP3OfRvcVArioeS0DaWN0NnL9eRIpKr61PrQCqowV-jMgoAtzBJskwk4V4x5vDQiSXeo7nKQ/s2910/Understanding-Adaptive-Machine-Translation_11.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1337" data-original-width="2910" height="147" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiOnlF3lWWnYxRD6d4jrbW1X4Z22Bihbt6KJUqt5H36RVH_pfA9OQGjDz4s22M6la1yI0WkpbGLuXoMVK8Q1C0TiD_2mLnp678nDoqwp1D_g-wq76_rhWKP3OfRvcVArioeS0DaWN0NnL9eRIpKr61PrQCqowV-jMgoAtzBJskwk4V4x5vDQiSXeo7nKQ/s320/Understanding-Adaptive-Machine-Translation_11.png" width="320" /></a></div><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">One should not assume that all adaptive MT systems follow the same technological approach. In fact, real-time, in-context adaptation can be achieved through various architectural methods. Upon closer examination of other adaptive MT solutions, it is apparent that dynamic adaptation can be accomplished using different technological strategies. However, as more buyers come to realize that the responsiveness of the MT system holds greater importance than a static COMET score on a random test set, the evaluation strategies will evolve. It will be more beneficial to assess which systems can adapt most effortlessly with minimal effort.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-weight: 700;">The ModernMT approach to adaptation is to bring the encoding and decoding phases of model deployment much closer together, allowing dynamic and active human-in-the-loop corrective feedback, that is not so different from the in-context corrections and prompt modifications we are seeing with large language models.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">It is possible that in the future, as Large Language Models (LLMs) become more cost-effective, scalable, secure, and controllable, they could be utilized to enhance SOTA adaptive MT models. This could improve both core translation quality and output fluency, either as stand-alone solutions or more likely as hybrid models that work with MT purpose-focused models that are yet to come.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">Although LLMs have demonstrated their effectiveness in certain high-resource languages, their performance in lower-resource languages is notably poor according to initial evaluations. This outcome is not surprising, given that LLMs are not specifically tailored for translation tasks. The challenge lies in the fact that LLMs rely on extensive data caches for each language, and the significant data volumes required for improvement are often difficult to locate. Therefore, resolving this issue will not be a quick or simple process.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">In contrast, ModernMT just <a data-mce-href="https://translated.com/adaptive-machine-translation-200-languages?ref=blog.modernmt.com" href="https://translated.com/adaptive-machine-translation-200-languages?ref=blog.modernmt.com" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #337ab7; text-decoration-line: none;">announced support for 200 languages</a> that can all immediately benefit from the continuous improvement infrastructure that underlies the technology, and begin the steady quality improvement process that is described in this article.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; color: #333333; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; font-size: 14px; margin: 0px 0px 10px;">It has become evident that real-time systems that can enhance performance and swiftly respond to informed feedback from experts are highly favored to tackle the task of large-scale automated language translation.</p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; box-sizing: border-box; font-family: "Helvetica Neue", Helvetica, Arial, sans-serif; margin: 0px 0px 10px;"><span style="color: #2b00fe; font-size: large;"><span style="background-color: white;"><b>Once customers have experienced the advantages of dynamic adaptive systems, they are unlikely to revert to the complexities, inconveniences, and slowly improving output quality of batch-trained static MT systems.</b></span></span></p><div class="separator" style="clear: both; text-align: center;"><br /></div><div class="separator" style="clear: both; text-align: left;"><br /></div></div>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0tag:blogger.com,1999:blog-6748877443699290050.post-79598938334398723762023-05-21T15:23:00.014-07:002023-05-26T15:59:39.950-07:00The Limits of Language AI <p><i><span style="color: #cc0000; font-family: trebuchet;">This is an article that I originally wrote about 18 months ago that was published in the <a href="https://imminent.translated.com/annual-report-2022">Imminent 2022 Annual Report,</a> long before ChatGPT was announced, and was revamped and updated to be published in the <a href="https://multilingual.com/issues/april-2023/the-limits-of-ai-with-language/">Multilingual Issue of April 2023</a>. </span></i></p><p><i><span style="color: #cc0000; font-family: trebuchet;">None of the main arguments have changed with the introduction of ChatGPT. All the structural problems identified with LLMs years ago are still present and have not been alleviated with the introduction of either ChatGPT or GPT-4. </span></i></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj4bB9u5hu_hRjPU7kZa6f4YzpLF8YJAiuN1UcDyIMEuckYyJDJ-ld7IPNpUEF5Ze3Ht1eWXZQrB0yTm9Gzuyqxaje4U4utDs1aTX1KSUuewVkQ9197uXfghAypDMKdeMEgsT7M5gsmSx9N8WJMvfgpNJPqF8ou1wn9ykSFU-zPcNFsbJcDsi9wru1Dcg/s1536/214_Vashe_Cover-1536x864.jpg" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="864" data-original-width="1536" height="180" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj4bB9u5hu_hRjPU7kZa6f4YzpLF8YJAiuN1UcDyIMEuckYyJDJ-ld7IPNpUEF5Ze3Ht1eWXZQrB0yTm9Gzuyqxaje4U4utDs1aTX1KSUuewVkQ9197uXfghAypDMKdeMEgsT7M5gsmSx9N8WJMvfgpNJPqF8ou1wn9ykSFU-zPcNFsbJcDsi9wru1Dcg/s320/214_Vashe_Cover-1536x864.jpg" width="320" /></a></div><br /><p><span style="font-family: inherit; font-size: 16px;">Large language models (LLMs) are all the rage nowadays and it is almost impossible to get away from the news frenzy around ChatGPT, BingGPT, and Bard. There is much talk about reaching artificial general intelligence, (AGI) but should we be worried that the machine will shortly take over all kinds of knowledge work, including translation? Is language really something that machines can master?</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Machine learning applications around natural language data have been in full swing for over five years now. In 2022 natural language processing (NLP) oriented research announced breakthroughs in multiple areas, but especially around improving neural machine translation (NMT) systems and neural language generating (NLG) systems like the Generative Pre-trained Transformer 3 (GPT-3) and ChatGPT, a chat-enabled variation of GPT-3 which can produce human-like, if inconsistent, digital text. It predicts the next word given a text history, and often the generated text is relevant and useful. This is because it has trained on billions of sentences and has the ability to often glean the most relevant material related to the prompt from the data it has seen.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">GPT-3 and other LLMs can generate algorithm-written text often nearly indistinguishable from human-written sentences, paragraphs, articles, short stories, and more. They can even generate software code that draws on troves of previously seen code examples. This suggests that these systems could be helpful in many text-heavy business applications and possibly enhance enterprise-to-customer interactions involving textual information in various forms.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">The original hype and excitement around GPT-3 have triggered multiple similar initiatives across the world, and we see today that the massive corpus of 175 billion parameters used in building GPT-3 has already been overshadowed by several other models that are even larger — Gopher from Deepmind has been built with 280 billion parameters and claims better performance in most benchmark tests used to evaluate the capabilities of these models. </span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">More recently, <a href="https://multilingual.com/tag/chatgpt" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; color: #2ea3f2; letter-spacing: inherit; line-height: 1em; margin: 0px; outline: 0px; padding: 0px 0px 10px; text-align: inherit; text-decoration-line: none; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">ChatGPT has taken the world by storm</a> and many knowledge workers are fearful of displacement from all the hype, even though we see the same problems with all LLMs, a lack of common sense, the absence of understanding, and the constant danger of misinformation and hallucinations. <span style="color: red;">Not to mention the complete disregard for data privacy and copyright in their creation.</span> </span></b></p><div style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b><span style="font-size: 16px;">GPT-4 parameter and training data overviews have been kept secret by OpenAI, which has now decided to cash in on the LLM gold rush, but many estimate parameters to be in the trillion-plus range. </span></b></span></div><div style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b><span style="font-size: 16px;">It is worth stating the originally stated OpenAI ethos has faded now, and one wonders if it was ever taken seriously. Their original mission statement was:</span></b> </span></div><blockquote><span style="color: #2b00fe; font-family: inherit;"><b>"Our goal is to advance digital intelligence in the
way that is most likely to benefit humanity as a whole, unconstrained by
a need to generate a financial return. Since our research is free from
financial obligations, we can better focus on a positive human impact." </b></span></blockquote><p></p><p style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit; font-size: medium;"><b>In an age where we face bullshit everywhere we turn, why should this be an exception?</b></span></p><div style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit; font-size: 16px;">The hype around some of these “breakthrough” capabilities inevitably raises questions about the increasing role of language AI capabilities in a growing range of knowledge work. Are we likely to see an increased presence of machines in human language-related work? Is there a possibility that machines can replace humans in a growing range of language-related work?</span></div><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">A current trend in LLM development is to design ever-larger models in an attempt to reach new heights, but no company has rigorously analyzed which variables affect the power of these models. These models can often produce amazing output but also have a fairly high level of completely wrong factual “hallucinations” where the machine simply pulls together random text elements in so confident a manner as to often seem valid to an untrained eye. But many critics are saying that larger models are unlikely to solve the problems that have been identified — namely, the textual fabrication of false facts, the absence of comprehension, and common sense.</span></p><blockquote style="font-size: x-large;"><b><span style="color: #2b00fe; font-family: inherit;">"ChatGPT “wrote” grammatically flawless but flaccid copy. It served up enough bogus search results to undermine my faith in those that seemed sound at first glance. It regurgitated bargain-bin speculations about the future of artificial intelligence. "</span></b></blockquote><blockquote><b><span style="color: #2b00fe; font-family: inherit;"><span style="text-align: right;">Trey Popp, </span><a href="https://thepenngazette.com/alien-minds-immaculate-bullshit-outstanding-questions/">Alien Minds, Immaculate Bullshit, Outstanding Questions</a></span></b></blockquote><p></p><div class="-trigger"><span style="font-family: inherit;"><br /></span></div><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">The initial euphoria is giving way to an awareness of the problems that are also inherent in LLMs and an understanding that adding more data and more computing power cannot and will not solve the toxicity and bias problems that have been uncovered. Critics are saying that scale does not seem to help much when it comes to “understanding,” and building GPT-4 with 100 trillion parameters, at a huge expense, may not help at all. The toxicity and bias that are inherent in these systems will not be easily overcome without strategies that involve more than simply adding more data and applying more computing cycles. However, what these strategies are, is not yet clear though many say this will require looking beyond machine learning.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">GPT-3 and other LLMs can be fooled into creating incorrect, racist, sexist, and biased content devoid of common sense. The model’s output depends on its input: garbage in, garbage out.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b><span style="color: #2b00fe; font-size: large;">"Just Calm Down About GPT-4 Already and stop confusing performance with competence. </span></b><b><span style="color: #2b00fe; font-size: large;">What the large language models are good at is saying what an answer should <em>sound like</em>, which is different from what an answer should <em>be</em>."</span></b></span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: right; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit; font-size: medium;"><a href="https://spectrum-ieee-org.cdn.ampproject.org/c/s/spectrum.ieee.org/amp/gpt-4-calm-down-2660261157"> Rodney Brooks</a><span> </span></span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Techniques like reinforced learning from human feedback (RLHF) can help to build guardrails against the most egregious errors but also reduce the scope of possible right answers. Many say this technique cannot solve all the problems that can emerge from algorithmically produced text as there are too many unpredictable scenarios.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">If you dig deeper, you discover although its output is grammatical, and even impressively idiomatic, its comprehension of the world is often seriously off. You can never really trust what it says. Unreliable AI that is pervasive and ubiquitous is a potential creator of societal problems on a grand scale.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Despite the occasional or even frequent ability to produce human-like outputs, ML algorithms are at their core only complex mathematical functions that map observations to outcomes. They can forecast patterns that they have previously seen and explicitly learned from. Therefore, they’re only as good as the data they train on and start to break down as real-world data starts to deviate from examples seen during training.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">In December 2021, an incident with Amazon Alexa exposed the problem that language AI products have. Alexa told a child to essentially electrocute herself (touch a live electrical plug with a penny) as part of a challenge game. This incident — and many others with LLMs — show that these algorithms lack comprehension and common sense, and can make nonsensical suggestions that could be dangerous or even life-threatening.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">“No current AI is remotely close to understanding the everyday physical or psychological world, what we have now is an approximation to intelligence, not the real thing, and as such it will never really be trustworthy,” said Gary Marcus in response.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Large pre-trained statistical models can do almost anything, at least enough for a proof of concept, but there is little they can do reliably because they skirt the required foundations.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Thus we see an increasing acknowledgment from the AI community that language is indeed a hard problem — one that cannot necessarily be solved by using more data and algorithms alone, and other strategies will need to be employed. This does not mean that these systems cannot be useful. Indeed, we understand how they are useful but have to be used with care and human oversight, at least until machines have more robust comprehension and common sense.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">We already see that machine translation (MT) today is ubiquitous, and by many estimates is responsible for 99.5% or more of all language translation done on the planet on any given day. But we also see that MT is used mostly to translate material that is voluminous, short-lived, transitory and that would never get translated if the machine were not available. Trillions of words a day are being translated by MT daily, yet when it matters, there is always human oversight on translation tasks that may have a high impact, or when there is a greater potential risk or liability from mistranslation.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">While machine learning use cases continue to expand dramatically, there is also an increasing awareness that a human-in-the-loop is often necessary since the machine lacks comprehension, cognition, and common sense.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit;">As Rodney Brooks, the co-founder of iRobot, said in a post entitled <em style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">An Inconvenient Truth About</em> AI: </span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit;"><br /></span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit; font-size: large;">“Just about every successful deployment of AI has either one of two expedients: It has a person somewhere in the loop, or the cost of failure, should the system blunder, is very low.”</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto;"><tbody><tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhTTKP2vjEuQ2mfaQqmYLp9kvps4J40jyodrpNhicdC-h016y2JJZS5KwtBo7shoPOHGjAPi_Lax14uaQkJu5dkrZc2QFUGoVu5hYVyeuih1UrxztuqCbcXQeSLk0qBqtVAUr_xahfxjgRutWe_nXegugZFIKF3wS3KBXwSByYg6u250BRrMKb4DCuZww/s1200/214_Vashee_Fig1.jpg" style="margin-left: auto; margin-right: auto;"><span style="font-family: inherit;"><img border="0" data-original-height="594" data-original-width="1200" height="198" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhTTKP2vjEuQ2mfaQqmYLp9kvps4J40jyodrpNhicdC-h016y2JJZS5KwtBo7shoPOHGjAPi_Lax14uaQkJu5dkrZc2QFUGoVu5hYVyeuih1UrxztuqCbcXQeSLk0qBqtVAUr_xahfxjgRutWe_nXegugZFIKF3wS3KBXwSByYg6u250BRrMKb4DCuZww/w400-h198/214_Vashee_Fig1.jpg" width="400" /></span></a></td></tr><tr><td class="tr-caption" style="text-align: center;"><span style="font-family: inherit; font-size: 14px;">Fig. 1. Linguistic communication of thoughts by speaker and listener</span></td></tr></tbody></table><span style="font-family: inherit;"><br /><br /></span><div style="text-align: left;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: large; font-weight: 700; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">What is it about human language that makes it a challenge for machine learning?</span></div><div style="text-align: left;"><span style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: large; font-weight: 700; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><br /></span></div><div><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Members from the singularity community summarized the problem quite neatly. They admit that “language is hard” when they explain why <a href="https://multilingual.com/issues/february-2023/the-great-gap-will-mt-ever-be-on-par-with-human-translators/" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; color: #2ea3f2; letter-spacing: inherit; line-height: 1em; margin: 0px; outline: 0px; padding: 0px 0px 10px; text-align: inherit; text-decoration-line: none; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">AI has not mastered translation yet</a>. Machines perform best in solving problems that have binary outcomes. </span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><a href="https://singularityhub.com/2018/03/04/why-hasnt-ai-mastered-language-translation/">Michael Housman, a faculty member of Singularity University,</a> explained that the ideal scenario for machine learning and artificial intelligence is something with fixed rules and a clear-cut measure of success or failure. He named chess as an obvious example and noted machines were able to beat the best human Go player. This happened faster than anyone anticipated because of the game’s very clear rules and limited set of moves. </span></p><div style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit; font-size: large;"><b><span style="color: #2b00fe;">Machine learning works best when there is one or a defined and limited set of correct answers</span>.</b></span></div><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Housman elaborated, <b>“Language is almost the opposite of that. There aren’t as clearly-cut and defined rules. The conversation can go in an infinite number of different directions</b>. And then, of course, you need labeled data. You need to tell the machine to do it right or wrong.”</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">Housman noted that it’s inherently difficult to assign these informative labels. “Two translators won’t even agree on whether it was translated properly or not,” he said. “Language is kind of the wild west, in terms of data.”</span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit; font-size: large;">Another issue is that language is surrounded by layers of situational and life context, intent, emotion, and feeling. The machine simply cannot extract all these elements from the words contained in a sentence or even by looking at hundreds of millions of sentences. </span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">The same sequence of words could have multiple different semantic implications. What lies between the words is what provides the more complete semantic perspective, and this is learning that machines cannot extract from a sentence. The proper training data to solve language simply does not exist and will likely never exist even though current models seem to have largely solved the syntax problem with increasing scale.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">Concerning GPT-3/4 and other LLMs: The trouble is that you have no way of knowing in advance which formulations will or won’t give you the right answer. GPT’s fundamental flaws remain. Its performance is unreliable, causal understanding is shaky, and incoherence is a constant companion.</span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Hopefully, we are now beginning to understand that adding more data does not solve the overall problem, even though it appears to have largely solved the syntax issue. More data makes for a better, more fluent approximation to language; it does not make for trustworthy intelligence.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">The claim to these systems are early representations of machine sentience or AGI is particularly problematic, and some critics are quite vocal in their criticism of these overreaching pronouncements and forecasts.</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><a href="https://www.technologyreview.com/2020/08/22/1007539/gpt3-openai-language-generator-artificial-intelligence-ai-opinion/">Summers-Stay said this about GPT-3:</a> “[It’s] odd, because it doesn’t ‘care’ about getting the right answer to a question you put to it. It’s more like an improv actor who is totally dedicated to their craft, never breaks character, and has never left home but only read about the world in books. Like such an actor, when he doesn’t know something, he will just fake it. You wouldn’t trust an improv actor playing a doctor to give you medical advice.”</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">Ian P. McCarthy said, “A liar is someone who is interested in the truth, knows it, and deliberately misrepresents it. In contrast, a bullshitter has no concern for the truth and does not know or care what is true or is not.” Gary Marcus and Ernest Davis characterize GPT-3 and new variants as “fluent spouters of bullshit” that even with all the data are not a reliable interpreter of the world.</span></b></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">For example, Alberto Romero says: “The truth is these systems aren’t masters of language. They’re nothing more than mindless ‘stochastic parrots.’ <b>They don’t understand a thing about what they say, and that makes them dangerous. They tend to ‘amplify biases and other issues in the training data’ and regurgitate what they’ve read before, but that doesn’t stop people from ascribing intentionality to their outputs.</b> GPT-3 should be recognized for what it is: a dumb — even if potent — language generator, and not as a machine so close to us in humanness as to call it ‘self-aware.’”</span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">The most compelling explanation that I have seen on why language is hard for machine learning is by Walid Saba, founder of Ontologik.Ai. <a href="https://thegradient.pub/machine-learning-wont-solve-the-natural-language-understanding-challenge/">Saba points out that</a> Kenneth Church, a pioneer in the use of empirical methods in NLP i.e. using data-driven, corpus-based, statistical, and machine learning (ML) methods was only interested in solving simple language tasks — the motivation was never to suggest that this technique could somehow unravel how language works, but rather he meant, “It is better to do something simple than nothing at all.” </span></p><p class="p1" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">However, subsequent generations misunderstood this empirical data-driven approach which was originally only intended to find practical solutions to simple tasks, to be a paradigm that will scale into full natural language understanding (NLU).</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">This has led to widespread interest in the development of LLMs and what he calls<b> “a futile attempt at trying to approximate the infinite object we call natural language by trying to memorize massive amounts of data.” </b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">While he sees some value in data-driven ML approaches for some NLP tasks (summarization, topic extraction, search, clustering, NER) he sees this approach as irrelevant for natural language understanding (NLU) where understanding requires a much more specific and accurate understanding of the one and only one thought that a speaker is trying to convey. Machine learning works on the specified NLP tasks above because they are consistent with the probably approximately correct (PAC) paradigm that underlies all machine learning approaches, but he insists that this is not the right approach for “understanding” and NLU.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">He explains that there are three reasons why NLU or "understanding" is so difficult for machine learning:</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span color="inherit" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: medium; font-weight: 700; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">1. The missing text phenomenon (MTP) is believed to be at the heart of all challenges in NLU.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">In human communication, <b>an utterance by a speaker has to be decoded to get to the specific meaning intended, by the listener, for understanding to occur.</b> There is often a reliance on common background knowledge so that communication utterances do not have to spell out all the context. That is, for effective communication, we do not say what we can assume we all know! This genius optimization process that humans have developed over 200,000 years of evolution works quite well, precisely because we all know what we all know. </span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">But this is where the problem is in NLU: machines don’t know what we leave out because they don’t know what we all know. The net result? NLU is difficult because <b>a software program cannot understand the thoughts behind our linguistic utterances if it cannot somehow “uncover” all that stuff that humans leave out</b> and implicitly assume in their linguistic communication. What we say is a fraction of all that we might have thought of before we speak.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><table align="center" cellpadding="0" cellspacing="0" class="tr-caption-container" style="margin-left: auto; margin-right: auto;"><tbody><tr><td style="text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiej0kdVAVgueuLAAAOcpWm1DtHCOPWIMJzltIxkNQDT-lI1DdOrDvz27QDBaiBUsa0tL2SV0XiGp0ERxRy7OsZOwkMAYLPgGKbVLXL4SCmo798vdYkZxxGjUXcpMN1xk10VEOYx7vnxRZoofL2G_uP-__4LqR67d3YrR2dGnIWUhnsDRhOoxgENtAH1A/s1536/214_Vashee_Fig2-1-1536x1081.jpg" style="margin-left: auto; margin-right: auto;"><span style="font-family: inherit;"><img border="0" data-original-height="1081" data-original-width="1536" height="281" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEiej0kdVAVgueuLAAAOcpWm1DtHCOPWIMJzltIxkNQDT-lI1DdOrDvz27QDBaiBUsa0tL2SV0XiGp0ERxRy7OsZOwkMAYLPgGKbVLXL4SCmo798vdYkZxxGjUXcpMN1xk10VEOYx7vnxRZoofL2G_uP-__4LqR67d3YrR2dGnIWUhnsDRhOoxgENtAH1A/w400-h281/214_Vashee_Fig2-1-1536x1081.jpg" width="400" /></span></a></td></tr><tr><td class="tr-caption" style="text-align: center;"><span style="font-family: inherit; font-size: 14px;">Fig. 2: Deep learning to create $30T in market cap value by 2037? (Source: ARK Invest).</span></td></tr></tbody></table><span style="font-family: inherit;"><br /></span><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span color="inherit" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: medium; font-weight: 700; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">2. ML approaches are not relevant to NLU: ML is compression, and language understanding requires uncompressing.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b><br style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; box-sizing: border-box; letter-spacing: inherit;" /></b>Our ordinary spoken language is highly (if not optimally) compressed. The challenge is in uncompressing (or uncovering) the missing text. Even in human communications, faulty uncompressing can lead to misunderstanding, and <b>machines do not have the visual, spatial, physical, societal, cultural, and historical context, all of which remain in the common understanding but unstated zone to enable understanding. This is also true to a lesser extent for written communication.</b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">What the above says is the following: machine learning is about discovering a generalization of lots of data into a single function. Natural language understanding, on the other hand, and due to MTP (missing text phenomena), requires intelligent “uncompressing” techniques that would uncover all the missing and implicitly assumed general knowledge text. Thus, he claims machine learning and language understanding are incompatible and contradictory. This is a problem that is not likely to be solved by 1000X more data and computing.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span color="inherit" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: medium; font-weight: 700; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">3. Statistical insignificance: ML is essentially a paradigm that is based on finding patterns (correlations) in the data.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Thus, the hope in that paradigm is that there are statistically significant differences to capture the various phenomena in natural language. Using larger data sets assumes that ML will capture all the variations. However, renowned cognitive scientist <b>George Miller said: “To capture all syntactic and semantic variations that an NLU system would require, the number of features [data] a neural network might need is more than the number of atoms in the universe!” </b>The moral here is this: Statistics cannot capture (nor even approximate) semantics even though increasing scale appears to have success with learning syntax. This fluency is what we see as "confidence" in the output.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><a href="https://www.thoughtco.com/pragmatics-language-1691654">Pragmatics studies how context contributes to meaning.</a> Pragmatist George Herbert Mead argued that communication is more than the words we use: “It involves the all-important social signs people make when they communicate.” <b>Now, how could an AI system access contextual information? It simply does not exist in the data they train on. </b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit; font-size: large;"><b><span style="color: #2b00fe;">The key issue is that ML systems are fed words (the tip of the iceberg), and these words don’t contain the necessary pragmatic information of common knowledge. </span><span style="color: red;">Humans can express more than words convey because we share a reality. But AI algorithms don’t.</span><span style="color: #2b00fe;"> AI is faced with the impossible task of imagining the shape and contours of the whole iceberg given only a few 2D pictures of the tip of the iceberg. </span></b></span></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7LqazcV2uWJAI_bPKwQrjKKd_mnsA0ai0JbynsbAZd49WxAGcH_aJXkNGpOKQEsVS3TD3asDYFunA1xX3ecvFJ_N7TqiwRC9o4J5lAoL4-1J8SRojpwimWqj0Es3Lk2xXaDeLSOJsgQjhLdiD7ZGEXdbEGj3bZh4cdZ3YO4ZjSVAuYjLgdpPFQ9nK2A/s960/Why%20Language%20is%20hard.png" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="225" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj7LqazcV2uWJAI_bPKwQrjKKd_mnsA0ai0JbynsbAZd49WxAGcH_aJXkNGpOKQEsVS3TD3asDYFunA1xX3ecvFJ_N7TqiwRC9o4J5lAoL4-1J8SRojpwimWqj0Es3Lk2xXaDeLSOJsgQjhLdiD7ZGEXdbEGj3bZh4cdZ3YO4ZjSVAuYjLgdpPFQ9nK2A/w400-h225/Why%20Language%20is%20hard.png" width="400" /></a></div><br /><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: center; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">Most of the data needed to achieve "understanding" is not available </p><span style="font-family: inherit;">Philosopher Hubert Dreyfus, a leading 20th-century critic, <a href="https://www.nature.com/articles/s41599-020-0494-4">argued against current approaches to AI</a>, saying that most of the human expertise comes in the form of tacit knowledge — experiential and intuitive knowledge that can’t be directly transmitted or codified and is thus inaccessible to machine learning. Language expertise is no different, and it’s precisely the pragmatic dimension often intertwined with tacit knowledge.</span><p></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="color: #2b00fe; font-family: inherit; font-size: medium;"><b>To summarize: we transmit highly compressed linguistic utterances that need a mind to interpret and “uncover” all the background information. This multi-modal, multi-contextual uncompression leads to “understanding</b>.” <b>Both communication and understanding require that humans fill in the unspoken and unwritten words needed to reach comprehension.</b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="color: #2b00fe; font-family: inherit; font-size: medium;"><b><br /></b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: center; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="color: #2b00fe; font-family: inherit; font-size: large;"><b>Languages are the external artifacts that we use to encode the infinite number of thoughts we might have. </b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: center; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="color: #2b00fe; font-family: inherit; font-size: large;"><b><br /></b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">In so many ways, then, in building ever-larger language models, <b>machine learning and data-driven approaches are trying to chase infinity in a futile attempt to find something that is not even “there” in the data.</b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Another criticism focuses more on the “general intelligence” claims being made about AI by people like OpenAI. Each of our AI techniques manages to replicate some aspects of what we know about human intelligence. But putting it all together and filling the gaps remains a major challenge. In his book, data scientist<a href="https://bdtechtalks.com/2021/03/29/ai-algorithms-representations-herbert-roitblat/"> Herbert Roitblat provides an in-depth review </a>of different branches of AI and describes why each of them falls short of the dream of creating general intelligence.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b>The common shortcoming across all AI algorithms is the need for predefined representations,</b> Roitblat asserts. Once we discover a problem and can represent it in a computable way, we can create AI algorithms that can solve it, often more efficiently than ourselves. It is, however, the undiscovered and unrepresentable problems that continue to elude us: the so-called edge cases. There are always problems outside the known set, and thus there are problems that models cannot solve.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">“These language models are significant achievements, but they are not general intelligence,” Roitblat said. <b>“Essentially, they model the sequence of words in a language. They are plagiarists with a layer of abstraction. Give it a prompt, and it will create a text that has the statistical properties of the pages it has read, but no relation to anything other than the language.</b> It solves a specific problem, like all current artificial intelligence applications. It is just what it is advertised to be — a language model. That’s not nothing, but it is not general intelligence.”</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">“Intelligent people can recognize the existence of a problem, define its nature, and represent it,” Roitblat writes. “They can recognize where knowledge is lacking and work to obtain that knowledge. <b>Although intelligent people benefit from structured instructions, they are also capable of seeking out their own sources of information.” </b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: center; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit; font-size: large;">In a sense, humans are optimized to solve unseen and new problems by acquiring and building the knowledge base needed to address these new problems.</span></b></p><h1 style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span color="inherit" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-weight: 700; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><br /></span></h1><div style="background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: left; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span color="inherit" style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: inherit; font-size: x-large; font-weight: 700; letter-spacing: inherit; margin: 0px; outline: 0px; padding: 0px; text-align: inherit; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;">The Path Forward</span></div><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Machine learning is being deployed across a wide range of industries, solving many narrowly focused problems and, when well implemented with relevant data, creating substantial economic value. This trend will likely only build momentum.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Some experts say that we are only at the beginning of a major value-creation cycle driven by machine learning that will have an impact as deep and as widespread as the development of the internet itself. The future giants of the world economy are likely to be companies that have leading-edge ML capabilities.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">However, we also know that AI lacks a theory of mind, common sense and causal reasoning, extrapolation capabilities, and a body, so it is far from being “better than us” at almost anything slightly complex or general. These are challenges that are not easily solved by deep learning approaches. We need to think differently and move on from more data plus more computing approaches to solving all our AI-related problems.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b style="font-size: 16px;">“The great irony of common sense — and indeed AI itself — is that it is stuff that pretty much <a href="http://kv-emptypages.blogspot.com/2021/01/adding-commonsense-reasoning-to-natural.html">everybody knows, yet nobody seems to know </a>what exactly it is or how to build machines that possess it,” said Gary Marcus, </b>CEO and founder of Robust.AI. “Solving this problem is, we would argue, the single most important step towards taking AI to the next level. Common sense is a critical component to building AIs that can understand what they read; that can control robots that can operate usefully and safely in the human environment; that can interact with human users in reasonable ways. <b><span style="color: #2b00fe;"><span style="font-size: medium;">Common sense is not just the hardest problem for AI; in the long run, it’s also the most important problem.</span>”</span></b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">Common sense has been called the <a href="https://www.newyorker.com/tech/annals-of-technology/can-computers-learn-common-sense">“dark matter of AI”</a> — both essential and frustratingly elusive. That’s because common sense consists of implicit information: the broad (and broadly shared) set of unwritten assumptions and rules of thumb that humans automatically use to make sense of the world. Critics of over-exuberant AI claims frequently point out that two-year children have more common sense than existing deep-learning-based AI systems whose “understanding” is often quite brittle.</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Common sense is easier to detect than to define, and its implicit nature is difficult to represent explicitly. Gary Marcus suggests <a href="https://www.noemamag.com/deep-learning-alone-isnt-getting-us-to-human-like-ai/">combining traditional AI approaches with deep learning</a>: “First, classical AI is a framework for building cognitive models of the world that you can then make inferences over. The second thing is, classical AI is perfectly comfortable with rules. It’s a strange sociology right now in deep learning where people want to avoid rules. They want to do everything with neural networks and do nothing with anything that looks like classical programming. But some problems are solved this way that nobody pays attention to, like making a Google Maps route.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">We need both approaches. The machine-learning stuff is pretty good at learning from data, but it’s poor at representing the kind of abstraction that computer programs represent. Classical AI is pretty good at abstraction, but it all has to be hand-coded, and there is too much knowledge in the world to manually input everything. So it seems evident that what we want is some kind of synthesis that blends these approaches.”</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">This is also the view of Yann LeCun (Head of AI, Meta) who won a Turing Prize and whose company has also released an open-source LLM. He does not believe that the current fine-tuning RLHF approaches can solve the quality problems we see today. The autoregressive models of today generate text by predicting the probability distribution of the next word in a sequence given the previous words in the sequence.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><b><a href="https://futurist.com/2023/02/13/metas-yann-lecun-thoughts-large-language-models-llms/">Autoregressive models are “reactive” and do not plan or reason,</a> according to LeCun. They make stuff up or retrieve stuff approximately, and this can be mitigated, but not fixed by human feedback. He sees LLMs as an “off-ramp” and not the destination of AI. LeCun has also said: “A system trained on language alone will never approximate human intelligence, even if trained from now until the heat death of the universe.”</b></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">LeCun has also proposed that one of the most important challenges in AI today is devising learning paradigms and architectures allowing machines to supervise their own world-model learning and then use them to predict, reason, and plan.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">Thus, when we consider the overarching human goal of understanding and wanting to be understood, we must admit that this is very likely always going to require a human in the loop, even when we get to building deep learning models with trillions of words. The most meaningful progress will be related to the value and extent of the assistive role that language AI will play in enhancing our ability to communicate, share, produce, and digest knowledge.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">Human-in-the-loop (HITL) is the process of leveraging machine power and enabling high-value human intelligence interactions to create continuously improving learning-based AI models. Active learning refers to humans handling low-confidence units and feeding improvements back into the model. Human-in-the-loop is broader, encompassing active learning approaches and data set creation through human labeling.</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit;">HITL describes the process when the machine is unable to solve a problem based on initial training data alone and needs human intervention to improve both the training and testing stages of building an algorithm. Properly done, this creates an active feedback loop allowing the algorithm to give continuously better results with ongoing use and feedback. With language translation, the critical training data is translation memory.</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">However, the truth is that there is no existing training data set (TM) so perfect, complete, and comprehensive as to produce an algorithm that consistently produces perfect translations.</span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-align: center; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="color: #2b00fe; font-family: inherit; font-size: large;">Again to quote Roitblat: “Like much of machine intelligence, the real genius [of deep learning] comes from how the system is designed, not from any autonomous intelligence of its own. Clever representations, including clever architecture, make clever machine intelligence.”</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;">This suggests that humans will remain at the center of complex, knowledge-based AI applications involving language even though the way humans work will continue to change. </span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><span style="font-family: inherit;">As the use of machine learning proliferates, there is an increasing awareness that humans working together with machines in an active-learning contribution mode can often outperform the possibilities of machines or humans alone. The future is more likely to be about how to make AI a useful assistant than it is about replacing humans.</span></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><span style="font-family: inherit;"><br /></span></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: "Noto Serif", Georgia, "Times New Roman", serif; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b>The contents of this blog are discussed in a live online interview format (with access to the recording) that Nimdzi shares on <a href="https://www.linkedin.com/events/thelimitsofaiwithlanguagefeat-k7062090703564099587/comments/">LinkedIn</a>. </b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: "Noto Serif", Georgia, "Times New Roman", serif; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"><b><i>The actual interview starts 13 minutes after the initial ads.</i></b></p><p style="--tw-border-spacing-x: 0; --tw-border-spacing-y: 0; --tw-ring-color: rgb(59 130 246 / 0.5); --tw-ring-offset-color: #fff; --tw-ring-offset-shadow: 0 0 #0000; --tw-ring-offset-width: 0px; --tw-ring-shadow: 0 0 #0000; --tw-rotate: 0; --tw-scale-x: 1; --tw-scale-y: 1; --tw-scroll-snap-strictness: proximity; --tw-shadow-colored: 0 0 #0000; --tw-shadow: 0 0 #0000; --tw-skew-x: 0; --tw-skew-y: 0; --tw-translate-x: 0; --tw-translate-y: 0; -webkit-font-smoothing: antialiased; background-attachment: initial; background-clip: initial; background-image: initial; background-origin: initial; background-position: initial; background-repeat: initial; background-size: initial; border-radius: 0px; border: none; box-shadow: none; box-sizing: border-box; font-family: "Noto Serif", Georgia, "Times New Roman", serif; font-size: 16px; margin: 0px; outline: 0px; padding: 0px 0px 1em; text-shadow: inherit; transition: none 0s ease 0s; vertical-align: baseline;"></p><div class="separator" style="clear: both; text-align: center;"><iframe allowfullscreen="" class="BLOG_video_class" height="266" src="https://www.youtube.com/embed/qRoqnloyg8M" width="320" youtube-src-id="qRoqnloyg8M"></iframe></div><br /><div class="separator" style="clear: both; text-align: center;"><br /></div><br /><b><br /></b><p></p></div><scribe-shadow id="crxjs-ext" style="height: 0px; left: 0px; overflow: visible; position: fixed; top: 0px; width: 0px; z-index: 2147483647;"></scribe-shadow>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com1tag:blogger.com,1999:blog-6748877443699290050.post-43992859718567973742023-04-14T13:28:00.000-07:002023-04-14T13:28:05.844-07:00The Challenge of MT with Non-English Language Pairs<p><span style="font-size: medium;"> <b>TL;DR Summary:</b><i> Evidence of superior performance from MT systems built by direct
modeling between languages and avoiding the use of pivoting through
English</i></span></p><p><br /></p><p>As the momentum for widespread Enterprise MT use continues, we are
increasingly seeing more interest in the use of MT across language pairs
that do not involve or include English. This can be a problem sometimes
as much of the long-term development of MT technology has been very
English-centric.</p><p><strong>Thus, MT most often works best in language combinations that go from or into English,</strong>
e.g., EN > FR, EN > DE, IT > EN, ZH > EN, or JP > EN. It
has also generally (though not strictly) been true that X > EN tends
to yield better results than EN > X. <strong>This is because there
simply is more data around English than any other language, both
bilingual text and especially monolingual text.</strong></p><p>But as
the global internet population changes, the importance of English
declines, and we see that there is growing interest in developing
optimized MT technology between languages that do not include English.
This is true across the globe as MT is seen as a technology that can
help to build momentum within regional trade zones.</p><p><strong>The need for translations between non-English pairs has typically been managed by going through English as a pivot language.</strong>
Thus, DE > IT would be accomplished in two steps, first using a DE
> EN system and then taking that output and using an EN > IT
system. This two-step journey also results in a degradation in output
quality as described below. This is beginning to change and there are an
increasing number of direct language combinations being developed
including the language pairs described in the Comparis case study
presented below.</p><p>It has often surprised me that some in the
translation industry use automated back translation as a way to check MT
quality, as from my vantage point it introduces similar problems to
what pivoting MT does. <strong>MT back translation by definition should
result in further deterioration of output quality as MT output will
often be something less than a perfect translation.</strong></p><p>This
point seems to evade many who advocate this method of evaluation, so let
us clarify with some mathematics as math is one of the few conceptual
frameworks available to man where the proof has meaningful certainty. If
one has a <strong>“perfect MT system”</strong> then the Source and Target segments should be very close to each other if not the same. <strong>The perfect score would be 1, which would mean that all the information in the source would be present in the target.</strong> Thus, anything less than perfect would be less than 1. So mathematically we could state this as:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh3IvgBiwox8W6XmrxLbivgTDh0L5phdr0YHarOVzrNoGCvu-U4he05uab2mbW6uQtU5wKWhrLFdrzonDUfXTjIl5kf4Feskd50crXDGXjEBXmwFbp-hONuKFNy2trkghpmDA5JTApSoCihugNWVEhrre124Fn27EPlMs9yyngBt66mua7iXqGcIYSBpw/s910/Perfect-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="495" data-original-width="910" height="174" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh3IvgBiwox8W6XmrxLbivgTDh0L5phdr0YHarOVzrNoGCvu-U4he05uab2mbW6uQtU5wKWhrLFdrzonDUfXTjIl5kf4Feskd50crXDGXjEBXmwFbp-hONuKFNy2trkghpmDA5JTApSoCihugNWVEhrre124Fn27EPlMs9yyngBt66mua7iXqGcIYSBpw/s320/Perfect-1.png" width="320" /></a></div>Thus, if we do a formal evaluation of the output of various MT systems <em>(each language direction should be considered a separate system) </em>and
find that the following table is true for our samples by running 5,000
sentences through various MT conversions (translations) and scoring each
MT translation (conversion) as a percentage “correct” in terms of
linguistic accuracy and precision.<div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgALt35EhtEKz4MhhGWEntrOxUndq4VSRNlS5LKJnpOpY85YsnbVAcItu1VOVUVU8lvNrRSJkxEruOGA20Gid-kUVkbRkfgbBAEMRtHCPnRQHy4b6npMtx2V_yZHm6KsV_pPgFFJJe2y7Y_G8ofoS5POxaG7_AyBxvqmV_I8sVF-lXLMLZUd0TRU9CJTg/s960/Accuracy-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="180" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgALt35EhtEKz4MhhGWEntrOxUndq4VSRNlS5LKJnpOpY85YsnbVAcItu1VOVUVU8lvNrRSJkxEruOGA20Gid-kUVkbRkfgbBAEMRtHCPnRQHy4b6npMtx2V_yZHm6KsV_pPgFFJJe2y7Y_G8ofoS5POxaG7_AyBxvqmV_I8sVF-lXLMLZUd0TRU9CJTg/s320/Accuracy-1.png" width="320" /></a></div><p>This evaluation gives us a sense of how these systems are likely to
perform with any new source material. But if we now chain the results
(to replicate the pivot effect) by making the output of one, the input
(source) of the other, we will find that results are different and get
continually worse e.g.</p><h3 id="es-en-de-85-x-70-0595-or-595-correct" style="text-align: center;"><strong>ES > EN > DE = .85 x .70 = 0.595 or 59.5% correct</strong></h3><h3 id="en-de-en-7-x-75-0525-or-525-correct" style="text-align: center;"><strong>EN > DE > EN = .7 x .75 = 0.525 or 52.5% correct</strong></h3><p>Of
course, the real results will not follow this kind of mathematical
precision and may actually be slightly better or worse. However, in
general, this degradation will hold. So now if we take our example and
run it through multiple iterations, we should expect to see a very
definite degradation of the output as we see below.</p><h3 id="en-es-en-from-mt-de-en-8-x-85-x-7-x-75-357" style="text-align: center;"><strong>EN > ES > EN (from MT) > DE > EN = .8 x .85 x .7 x .75 = 35.7%</strong></h3><p>This
is exactly the strategy that has been used by content creators in what
can be called MT-based humor. The translation degradation works even
better if multiple iterations are done with a larger variety of
languages. Even more so when the languages are very different
linguistically.</p><p><a href="https://www.youtube.com/watch?v=Rte07sl3Vf4&ref=blog.modernmt.com">Here is an example</a> where you can see the source at the top of the video and the multi-pivoted translation at the bottom.</p><div class="separator" style="clear: both; text-align: center;"><iframe allowfullscreen="" class="BLOG_video_class" height="266" src="https://www.youtube.com/embed/Rte07sl3Vf4" width="320" youtube-src-id="Rte07sl3Vf4"></iframe></div><br /><p>Here is <a href="https://www.youtube.com/watch?v=jN_qt5QlKQU&ref=blog.modernmt.com">a more recent example</a> that shows that even all the recent advances in NMT have not managed to reduce the degradation much if enough pivoting is done.</p><p>So,
while it is clear that simple two-step pivoting is not as damaging, it
does suggest that it is worth enabling direct translations between
non-English language combinations whenever possible.</p><p>The following
case study that describes the Comparis experience is helpful to
understand the benefit of avoiding pivoting whenever possible and
provides insight into some of the issues that came up in a comparative
evaluation.</p><p><br /></p><div><h1 style="text-align: left;"><strong>The Comparis Case Study </strong></h1><h2 style="text-align: left;"><strong>(written by Daniele Giulianelli)</strong></h2><p>Comparis is the #1 consumer empowerment platform in Switzerland: we
help our users find the products that most fit their needs in different
fields: health insurance, car insurance, mortgage, consumer finance, and
many more. Since Switzerland is “multilingual by design” (it has 4
national languages), translation and localization have a very important
role in our company.</p><p>Our content is mostly written in German, and
then gets translated into French and Italian (in their Swiss locale),
and into English:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgq9EcBTqcUubECDRgpS1aMvHRsZ3fHoN2bDLCDEnXZe7BJXAUvxpPsErK4mn5uGNjD5wEVGXM8yeSqAmXQVB7nH6ZFZsWnznJ-g3jUOHiIZIuI23VIB83xVyiuI5uECdAi50Ifnf3ayN2EY5E-dLyfSdubzCWDmE8Ke99UlzdGkI9acW5JtiXELy0nVg/s725/ModernMT-comparison-02-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="388" data-original-width="725" height="171" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgq9EcBTqcUubECDRgpS1aMvHRsZ3fHoN2bDLCDEnXZe7BJXAUvxpPsErK4mn5uGNjD5wEVGXM8yeSqAmXQVB7nH6ZFZsWnznJ-g3jUOHiIZIuI23VIB83xVyiuI5uECdAi50Ifnf3ayN2EY5E-dLyfSdubzCWDmE8Ke99UlzdGkI9acW5JtiXELy0nVg/s320/ModernMT-comparison-02-1.png" width="320" /></a></div><p>When starting a Machine Translation program, we had to face two main
issues: very in-domain content and small locales. Thus, generic MT was
not really an option, since the training data sets mostly refer to the
“bigger” locales (Germany, France, Italy). This would have resulted in a
huge post-editing effort and probably no real efficiency gain from the
use of MT.</p><p>The path to customization is not always easy. After
talking to some providers, it became clear that with our Translation
Memories (about 150,000 segments per language pair) it would have been
hard to have a significant impact on the quality of the MT output. Also,
a traditional “bulk” customization can be pretty expensive. <strong>And in most cases, there is no customization option for non-English language pairs, which are by far the most important for us.</strong></p><p>That’s how we chose to try ModernMT since it gave us the opportunity to:</p><ul><li>Customize “on the fly” by just adding TMX files.</li><li>Customize directly between non-English language pairs without pivoting.</li></ul><p>We
evaluated different MT solutions using ModelFront: the first candidates
were Google Translate, DeepL, and ModernMT. For the latter, we just
used one year (2020) of Translation Memory as a training data set for
customization. ModernMT proved to be the best solution, even with this
minimal customization:</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjwGze446-ahD8Xozl3oaQVuLa3zjhywUANg_1I1xpV1wc8vbgsCPCQzN_vuyI7LLNy0NmcToyRiKJIFXgskQz5Sf5oQqW6GIrW93-B5XYKaqMyDG_1CEhmAz8BWcqlQj0h_sBDJ0j2f_fkmy4gQbhd60kDt7tTndX1fMvcjzJdwKa2vX_LB8eYdh7i1Q/s854/ModernMT-comparison-01-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="405" data-original-width="854" height="152" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjwGze446-ahD8Xozl3oaQVuLa3zjhywUANg_1I1xpV1wc8vbgsCPCQzN_vuyI7LLNy0NmcToyRiKJIFXgskQz5Sf5oQqW6GIrW93-B5XYKaqMyDG_1CEhmAz8BWcqlQj0h_sBDJ0j2f_fkmy4gQbhd60kDt7tTndX1fMvcjzJdwKa2vX_LB8eYdh7i1Q/s320/ModernMT-comparison-01-1.png" width="320" /></a></div><div><br /></div>Here are a couple of examples to illustrate how ModernMT performs in our locales:</div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhf58qLISe6n3dlklgLtE7a_PIffqxeLy37b1AC7lx8wi2b_cJKuibMczAIuElUvCuqTPZUkCVznTjlyiEvUCwdDmActUOXBJg8SsX2LqbZPMMAMqa8_AY1J0uQeCJV2lh0Q0b5-4ibhp29OZbO4YLHDhcuZ18xPD1_X0_T7mApFGrdfR1dt4lRTc-iMQ/s960/Comparis-Txt-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="180" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhf58qLISe6n3dlklgLtE7a_PIffqxeLy37b1AC7lx8wi2b_cJKuibMczAIuElUvCuqTPZUkCVznTjlyiEvUCwdDmActUOXBJg8SsX2LqbZPMMAMqa8_AY1J0uQeCJV2lh0Q0b5-4ibhp29OZbO4YLHDhcuZ18xPD1_X0_T7mApFGrdfR1dt4lRTc-iMQ/s320/Comparis-Txt-1.png" width="320" /></a></div>Customizing with bilingual translation memory files (TMXs) without
pivoting also proved to be very efficient for MT performance on
in-domain topics:<div><br /><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgiWT1y9A1e-waCfiTfEdGzWYE332OyOJfw2yCEwNPJeK7_igdZ8YxDdoCLj5ljrvXCb0_73YAVIq61bcdoIWfiUSmdEps-Sbm0M14bhzotzMdMeCY2iKcoV41wxYFyblwZppaFn-sPdMLAneeOxLej4ihg8vX68ApWMzSwIaDMC2DYCbvOWwkHCqMJCA/s960/Comparis-Txt-2.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="180" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgiWT1y9A1e-waCfiTfEdGzWYE332OyOJfw2yCEwNPJeK7_igdZ8YxDdoCLj5ljrvXCb0_73YAVIq61bcdoIWfiUSmdEps-Sbm0M14bhzotzMdMeCY2iKcoV41wxYFyblwZppaFn-sPdMLAneeOxLej4ihg8vX68ApWMzSwIaDMC2DYCbvOWwkHCqMJCA/s320/Comparis-Txt-2.png" width="320" /></a></div><p>After analyzing these results, we decided to move forward with
ModernMT. Another reason for choosing this provider has been of course
their <strong>adaptive</strong> technology. Right from the start we noticed that our machine translation <strong>learns </strong>our terminology and our style while we post-edit the MT output.</p><p>We
were even able to solve our peculiar issue with the tone of voice for
Swiss Italian: which is neither formal nor informal. Just impersonal -
so instead of writing “you should compare insurances” we have “it is
important to compare insurances”. It took just a couple of weeks to have
it function in our MT output as we wanted.</p><p>All things considered, we observed almost a <strong>30% productivity boost </strong>in
our team across all languages. And we are confident that this trend
will continue, and that our post-editing effort will diminish. This will
give us more resources in our team to work on non-MT tasks like
advertising, slogans, and SEO optimization.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEikIPSz0ow_V3kwtDrNFyo_ZjmlsPo0v212ryXPstXtTf5vRiL2mFNboYT1errSBaXUFh_XCe3aTCYqatvpy9rOhfnHuBM5knyOzmkrMdfj-pSeA2FK1d0ix0Us0OWBTnQsa9FK3BxwIrS2ouddXn9kwSV5Bzi2xFPJuiMgtE8R3TybsP1qo0rY8hMBug/s960/Comparis-Txt-3.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="540" data-original-width="960" height="180" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEikIPSz0ow_V3kwtDrNFyo_ZjmlsPo0v212ryXPstXtTf5vRiL2mFNboYT1errSBaXUFh_XCe3aTCYqatvpy9rOhfnHuBM5knyOzmkrMdfj-pSeA2FK1d0ix0Us0OWBTnQsa9FK3BxwIrS2ouddXn9kwSV5Bzi2xFPJuiMgtE8R3TybsP1qo0rY8hMBug/s320/Comparis-Txt-3.png" width="320" /></a></div><br /><p><b><i><br /></i></b></p><p><b><i>Daniele Giulianelli has been in the localization industry for over 10
years, focused on finance and insurance. He is currently the Leader of
Translations and Product Owner Newsroom at Comparis, working to
establish best practices for the challenges of in-house language
services, especially focusing on MT and multilingual SEO.</i></b></p><div><br /></div><div><br /><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgVwemOuUsLQAh_pBzFbermki4huqblP-Yx7QpW6M8srPRhPdr3XlBwUzq1-7vDsNH9eNFfpp80oTDrDGhBc3utvFuYJqSUZ3bzxsPloGGDiKPhZDzaVV6-hzdaApjSIygJk4QTSqVnj6vGYZCfthpdkjjijrH6jEWqZutV80infDgR3AVAo1UTSkq6lQ/s685/Daniele-G.jpeg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="591" data-original-width="685" height="276" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEgVwemOuUsLQAh_pBzFbermki4huqblP-Yx7QpW6M8srPRhPdr3XlBwUzq1-7vDsNH9eNFfpp80oTDrDGhBc3utvFuYJqSUZ3bzxsPloGGDiKPhZDzaVV6-hzdaApjSIygJk4QTSqVnj6vGYZCfthpdkjjijrH6jEWqZutV80infDgR3AVAo1UTSkq6lQ/s320/Daniele-G.jpeg" width="320" /></a></div><br /><p><br /></p></div></div>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0tag:blogger.com,1999:blog-6748877443699290050.post-15313031965035152222023-02-17T12:27:00.015-08:002023-02-18T12:41:02.482-08:00The Problem With LangOps<p><span style="color: #660000;"><span style="font-family: georgia;"><i>This is a letter I wrote to the editor of Multilingual after reading several articles </i></span><i style="font-family: georgia;">focused on LangOps</i><i style="font-family: georgia;"> in <a href="https://multilingual.com/issues/december-2022/">the December 2022 issue</a>. This discussion <a href="https://www.linkedin.com/posts/multilingual-media_on-the-origin-of-langops-the-evolution-of-activity-7016494141467463680--08w/">started on LinkedIn</a> and Cameron invited the active contributors to formalize our comments and write a letter to the editor with alternate viewpoints.</i></span></p><p><i style="font-family: georgia;"><b><span style="color: #660000;">TLDR: LangOps is a term that refers to the vague use of "A.I." in/around localization or is nothing more than a way to describe the centralization of enterprise translation production processes.</span></b></i></p><p><i style="font-family: georgia;"><br /></i></p><p><span style="font-family: georgia;"></span></p><div class="separator" style="clear: both; text-align: center;"><span style="font-family: georgia;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEijLO75K5Ts0YVBNwrYcOUo4OM1bRqotpLmfQ2zBGW906XU4A1jF8aO-8_BVPQ19azCIXai6pFuux9ZhkmOxIuntY-XETUnI6kehADLERaoblKOxakPbL-XjiTfqODmHIaXDUceJlfsI_zszBFEvG4GCG8-M-4HlVOC_ZTBqhdCvG8up7vlEcXVFldjog/s800/LangOps.jpeg" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="450" data-original-width="800" height="225" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEijLO75K5Ts0YVBNwrYcOUo4OM1bRqotpLmfQ2zBGW906XU4A1jF8aO-8_BVPQ19azCIXai6pFuux9ZhkmOxIuntY-XETUnI6kehADLERaoblKOxakPbL-XjiTfqODmHIaXDUceJlfsI_zszBFEvG4GCG8-M-4HlVOC_ZTBqhdCvG8up7vlEcXVFldjog/w400-h225/LangOps.jpeg" width="400" /></a></span></div><span style="font-family: georgia;"><br /><i><span style="color: #660000;">I carefully read all of the following <b>before</b> writing my letter to ensure that I had not somehow missed the boat. The basic question I am still left with after looking carefully through the LangOps material is "Where's the real substance of this concept/idea/word?"</span></i><i><br /></i></span><p></p><p></p><ul style="text-align: left;"><li> Article by Renato: <a href="https://multilingual.com/issues/december-2022/langops-the-vision-and-the-reality/">LangOps- The Vision & the Reality</a> </li><li>Article by Arthur Weitzel: LangOps: <a data-mce-href="https://draft.blogger.com/blog/post/edit/6748877443699290050/1531303196503515222#" href="https://multilingual.com/issues/december-2022/langops-pipe-dream-lsps-heaven-or-just-a-new-hashtag/">Pipe Dream, LSP´s Heaven or Just a New Hashtag?</a></li><li>Article by Andrew Warner: <a href="https://multilingual.com/issues/december-2022/on-the-origin-of-langops-the-evolution-of-the-localization-roadmap/">On the Origin of LangOps</a> <strong>- </strong>The evolution of the localization roadmap</li><li>Article by Miguel Cerna: <a href="https://www.linkedin.com/pulse/langsops-localization-integration-rather-than-dr-miguel-cerna-%E9%A9%AC%E5%85%8B%E7%BD%97/">LangOps and Localization</a> Integration rather than substitution?</li><li>Article by Riteba McCallum: <a data-mce-href="https://multilingual.com/?p=193963" href="https://multilingual.com/?p=193963">The LangOps Paradigm: Perceptions of machine translation within the translation industr</a>y </li><li>and of course the LangOps principles.</li></ul><p><br /></p><div class="separator" style="clear: both; text-align: center;"><iframe allowfullscreen="" class="BLOG_video_class" height="266" src="https://www.youtube.com/embed/K6a-FLKMxUs" width="320" youtube-src-id="K6a-FLKMxUs"></iframe></div><div style="text-align: center;"><span style="color: #2b00fe;">Jump to 2' 50" to get to the relevant part</span></div><p><i><span style="color: #660000;">Here is a slightly ornamented version of <a href="https://multilingual.com/issues/february-2023/langops-wheres-the-beef/">the text of my letter to the editor which was published in the Multilingual February 2023 issue.</a> I include a version with emphasis (mine) so that others may also comment on this, and perhaps correct my misperception. </span></i></p><p><i><span style="color: #660000;">Special Thanks to <a href="https://www.linkedin.com/in/marjolein-groot-nibbelink/">Marjolein Groot Nibbelink </a>for taking the trouble to convert the letter to <a href="listen: https://multilingual.com/?p=198946">a really well-read audio track that can be played back faster.</a></span></i></p><p><i><span style="color: #660000;"><a href="listen: https://multilingual.com/?p=198946"><iframe height="150" src="https://609ac11b002ce7-92070854.castos.com/player/1415254" width="100%"></iframe>
</a></span></i></p><p><br /></p><p>Dear Multilingual Editor (Cameron),</p><p>After reading the various articles on
LangOps in the Multilingual December 2022 issue, I had hoped that I
would get a better sense of what LangOps is, and why it matters. But I
cannot say that this happened for me, and <b>I am not sure if I (or any
other reader) have any more clarity on what LangOps is, beyond it being a
vendor buzzword, </b>that remains fuzzy and amorphous because there is not
enough supporting evidence to document it properly. While there was much
discussion about why a new definition that went further than
localization is needed, there was not much that defined LangOps in more
concrete terms. <b>I suspect the fuzziness and lack of clarity that I felt
are true for many other readers as well.</b></p><h3 style="text-align: center;"><strong><span style="color: #2b00fe; font-size: large;">One is left asking. “Where’s the beef?” on this thing they call LangOps.</span></strong></h3><p>I
reviewed the articles in the magazine on the LangOps subject again
before writing this letter, to better identify the defining elements,
and to make sure I was fair and had not missed some obvious facts. My
intention with my comments here is to hopefully provide a coherent
critique of the subject matter, which started in discussion with comments made by several readers about LangOps <a href="https://www.linkedin.com/posts/multilingual-media_on-the-origin-of-langops-the-evolution-of-activity-7016494141467463680--08w/">on LinkedIn.</a> </p><p>From my
reading, the articles in Multilingual were clearer on <b>Why</b> new
definitions are needed, but <b>less clear on the What [it is] or explaining the How.</b></p><p>It
appears to me that the LangOps concept is another attempt by some
stakeholders in the industry to raise the profile of the translation
business, to make it more visible at the executive level, or to increase
the perceived value of the translation production process by imbuing it
with more complexity and mysterious undefined AI elements. However, in
the absence of specifics, it becomes just another empty buzzword that
creates more confusion than clarity for most of us, especially so for
new buyers. </p><p><b>It is difficult to see how any sponsor could take the
descriptions provided in this issue of Multilingual to a senior executive to ask for
funding, or even to explain what it is.</b></p><p>It is clear that as the
translation of some product and marketing content became recognized as a
valuable international business-driving activity, the need to scale,
organize and systematize it became more urgent and led to what most call
localization today. </p><p><b>Thus, localization I think refers to the many
processes, activities, and tools, used in making language translation
processes more automated, structured, and systematic.</b> Most often this
work is related to relatively static content that is mandatory in
international markets, but recently it has expanded to include more
customer service and support content. It also sometimes includes
cultural adaptations that are made in addition to the basic
translation. </p><p>TMS systems have been central to the localization worldview
over the past decade, as these TMS systems facilitate the development
and management of different workflows, monitor translation work, and
ease project management of distributed translation-related tasks (TEP).
It is also true that MT has been minimally used in hard-core
localization settings as MT systems were not deemed to be accurate,
flexible, and simple enough to configure to be used in this work.</p><p>By carefully reviewing the published Multilingual articles again, I
gathered that the following elements that are being used to define what
LangOps is:</p><ul><li>There are AI-driven capabilities applied to certain localization processes <b>which are not defined,</b></li><li><strong>Centralization of all translation production activities across the enterprise,</strong></li><li>Introduction of “more” technology into existing localization workflows, <b>but what this is specifically, is unclear,</b></li><li>LangOps is said to be made up of cross-functional and inter-disciplinary teams, <b>but who and why is not clear,</b></li><li><strong>Possibly </strong>adding other value-added language tasks (sentiment analysis, summarization, chatbots) in addition to the translation. [This at least is clear].</li></ul><p style="text-align: center;"><span style="color: #2b00fe; font-size: large;"><strong>To my view, the only element here that is clear in the many descriptions [of LangOps] is that of the centralization of translation production.</strong> </span></p><p>The other elements used to describe what it is are kind of fuzzy and
hard to pin down. They can mean anything or could mean nothing since
vagueness is not easily pinned down. LangOps is another term, that is
possibly even worse than localization <em>(which confuses many regular people and many new customers)</em>
because it creates a communication problem. </p><p>How do you answer the
question, “What do you do?” in an elevator, a cab, at a party, on an
airplane, with family and friends? As you can see both Localization and
LangOps present opaque, obfuscating images to the regular human mind.</p><p><b><span style="color: #2b00fe;">Would
it not be so much easier to just say “Language Translation to Drive
International Business”? And then maybe add, “We use technology, tools,
and people to do it at a large scale efficiently.”</span></b></p><p>I would like to
suggest a different way to view the continuing evolution of business
translation. It is my feeling that the LangOps movement is
linking the growing number of MT use cases, which have more dynamic IT
connectivity, and cross-organization collaboration implications, with a
need for a new definition.</p><p>We have now reached that perfect storm
moment where most B2C and B2B businesses recognize that they need a
substantial digital presence, that it is important to provide large
volumes of relevant content to serve and please their customers, and
that they need to listen to customers in social media, understand trends
faster, and communicate across the globe much faster. </p><p>This means that
successful businesses have to share, communicate, listen, and produce
translations at a much larger scale than they have had to in the past.
The core competency from traditional localization work is less likely to be
useful with these new challenges. These new market requirements need a
shift away from TM and TMS-managed work to a more MT-centric view of the
world. The volume of translation increases from thousands of translated
words, a month, to millions or even billions of words a month to drive
successful international business outcomes in the modern era. </p><p>As
Generative AI improves and begins to be deployed in production customer
settings, we will only see the translation volumes grow another 10X or
100X. Thus, deep MT competence increasingly becomes a core requirement
to be in the enterprise translation business.</p><p>MT has been improving dramatically over the last five years in
particular, and it is not ridiculous to say that it is getting close to
human output in some special cases when systems are properly designed
and deployed by competent experts. </p><p><b>Competence means that experts can
quickly adapt and modify MT systems to produce useful output in the
20-30 different use cases where an enterprise faces an avalanche of text
and/or audiovisual content.</b> The new use cases go beyond the traditional
focus of localization in terms of content and process. We now need to
translate much more dynamic content related to customer services and
support, translate more active communications (chat, email, forums),
share more structured and unstructured content, pay more attention to
social media feedback, and are just more real-time and dynamic in
general.</p><p>The successful modern global enterprise listens,
understands, communicates, and actively shares content across the globe
to improve customer experience. <strong>Thus, I think it is fair to say
that we (the translation business) are moving to a more MT-centric world
from a previously TMS-centric world, and a critical skill needed today
is deep competence with MT.</strong> </p><p><b>Useful MT output means it helps
grow and drive international business, even though it may not be
linguistically “perfect”.</b> The requirement for MT competence requires
moving far beyond choosing an MT system with the best BLEU or COMET
score. </p><p>MT Competence means you can find egregious errors <em>(MT & AI make these errors all the time)</em>
and instantly correct these problems to minimize damage. </p><p>MT Competence
means the skill and agility to respond to changing business needs and
new content types and the ability to rapidly modify MT systems as
needed. </p><p><span style="color: #2b00fe;"><b>Competence in managing rapid, responsive, deep adaptation of MT
systems will be a key requirement to actively participate as an
enterprise partner (not vendor) on a global stage very shortly.</b> </span></p><p>When
language translation is mission-critical and pervasive, the service
provider will likely evolve from being a vendor to being a partner. It
can also often mean that the scope of localization teams is greatly
expanded and become more mission-critical.</p><p>While I can see a
business reality where there is Machine-First & Human Optimized translation approach to content across the global enterprise, which requires responsive,
continuously improving MT, it also means moving beyond traditional MTPE
where clean-up crews come to reluctantly fix badly formed MT output
produced by inexperienced and incompetent MT practitioners. </p><p>However, the
lights start to dim for me when I think of "LangOps" being part of this
reality in any form whatsoever.</p><p>This continuing evolution of
business translation also probably means that there is a much more
limited role for the TMS or using it only for some localization
(software and key documentation) workflows. The more common case as
translation volumes grows is to connect all (Customer Experience)
CX-related text directly into highly tuned, carefully adapted NMT
systems in high-performance low-latency IT infrastructure that is directly customer-facing, or customer accessible. </p><p><strong>Recent data I have seen on MT use across a broad swathe of enterprise users shows that as much as 95% of MT use completely bypasses the TMS</strong>.
Properly tuned expert-built MT engines do not need the unnecessary
overhead of a TMS system. The enterprise objective is to enable
translation at scale for everything that might require instant, and
mostly but not necessarily a perfectly accurate translation, as long as
it furthers and enhances any and every global business initiative and
communication. </p><p><strong>Speed and scale are more important and have a
more positive impact on international business success in many
CX-related use cases than perfect linguistic quality does.</strong> The enterprise executives understand this even though we as an industry might not.</p><p><strong>I am not aware of a single LangOps configuration or group on this earth</strong>
or know any enterprise that claims to have such an initiative, but I
can point to several massive-scale MT-driven translation engines around
the world e.g. Airbnb, Amazon, Alibaba, and eBay where billions of words
are translated regularly to drive international business and customer
delight and serve a growing international customer base. I am confident
we will see this pool of enterprise users grow beyond the eCommerce
markets.</p><p><b><span style="color: #2b00fe;">Thus, I see little value in promoting the concept of
LangOps as what actually seems to be happening is that more expert-tuned
enterprise MT is being used and we see the share of MT used to total
translation volumes continue to grow. </span></b></p><p>As this kind of responsive, highly
adaptive MT capability becomes more pervasive across an enterprise, it
also becomes a critical requirement for international business success.
The activities related to organizing and managing significantly more
dynamic content and translation volumes should not be mistaken to be
something as vague as LangOps, as no organization I am aware of has the
building blocks or template to create such a vaguely defined function. I think that it
is more likely that Localization teams will evolve and the scope of
their activities will increase, perhaps as dramatically as we have seen at
Airbnb.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj5p5Chr_2GnRCq9WOdHKRENppJrH2ldSx6Q7pR9MrkvOQytXbKTOFSP6oWIw9ldj1a5e3X3H97TgHWXK7hdcZe5Yehri_vP7WEpGL2RUW4h0QJQIWMBFRvwtdLydMH8B57ysNsDjm7ZyljiO52sL3VOcooTld9_i6-9HfQ_-QJnzo8LOi_D87bbzJeVw/s1198/AIrbnb%20Revenue.jpg" style="margin-left: 1em; margin-right: 1em;"><img alt="Airbnb just booked its first annual profit in its near-15-year history, a whopping $1.9bn in 2022. It now appears to be in rarefied air, with its place as the de facto online marketplace for homestays and experiences, giving it a network effect that’s hard to compete with." border="0" data-original-height="1198" data-original-width="1198" height="400" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj5p5Chr_2GnRCq9WOdHKRENppJrH2ldSx6Q7pR9MrkvOQytXbKTOFSP6oWIw9ldj1a5e3X3H97TgHWXK7hdcZe5Yehri_vP7WEpGL2RUW4h0QJQIWMBFRvwtdLydMH8B57ysNsDjm7ZyljiO52sL3VOcooTld9_i6-9HfQ_-QJnzo8LOi_D87bbzJeVw/w400-h400/AIrbnb%20Revenue.jpg" width="400" /></a></div><p style="clear: both; text-align: center;"><span style="font-size: x-small;">Airbnb just booked its first annual profit in its near-15-year history, a
whopping $1.9bn in 2022. It now appears to be in rarefied air, with<strong data-stringify-type="bold"> its place as the de facto online marketplace for homestays and experiences,</strong> giving it a network effect that’s hard to compete with.</span></p><p style="clear: both; text-align: left;"><span style="color: #1d1c1d; font-size: x-small;"><span style="background-color: #f8f8f8; font-variant-ligatures: common-ligatures;"> </span></span></p><p>I did find all the articles on LangOps useful in
furthering my understanding, especially the ones by Riteba McCallum, and
Miguel Cerna, and my comments should not be mistaken as a wholesale
dismissal of the viewpoints presented. On the contrary, I think we have
much more agreement on many of the core issues discussed. Though I do
admit that I find the general concept of LangOps as it has been painted,
to be a likely hindrance to our mutual future rather than a beneficial
concept to drive our success with globalization and international
business initiatives with our common customers.</p><p><br /></p><p>Respectfully Yours,</p><p>Kirti Vashee</p><p><br /></p><p><br /></p><p>Here is the LinkedIn article where the discussion began:
<iframe allowfullscreen="" frameborder="0" height="570" src="https://www.linkedin.com/embed/feed/update/urn:li:share:7016494140750245888" title="Embedded post" width="504"></iframe>
</p><p><br /></p><p>P.S. Maybe all I am saying is that LangOps just needs more cowbell 😄😄😄 to get the sound and the concept right?</p><div class="separator" style="clear: both; text-align: center;"><iframe allowfullscreen="" class="BLOG_video_class" height="266" src="https://www.youtube.com/embed/cVsQLlk-T0s" width="320" youtube-src-id="cVsQLlk-T0s"></iframe></div><br /><p><br /></p>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com14tag:blogger.com,1999:blog-6748877443699290050.post-6489066427577278972023-02-01T16:40:00.001-08:002023-02-01T16:40:29.653-08:00The March Towards AI Singularity and Why It Matters<p> </p><h3 id="why-progress-in-mt-is-a-good-proxy-of-progress-with-the-technological-singularity">Why progress in MT is a good proxy of progress with the technological singularity</h3><div><br /></div><div>For as long as machine technology has been around (now over 70 years)
there have been regular claims made by developers of the technology
reaching “human equivalence”. <strong>However, until today we have not
had a claim that has satisfied practitioners in the professional
translation industry, who are arguably the most knowledgeable critics
around</strong>. For these users, the actual experience with MT has not
been matched by the many extravagant claims made by MT developers over
the years.</div><div><br /></div><div><strong><span style="color: #2b00fe;">This changes with the long-term study and translation production
data presented by Translated SRL at the AMTA conference which provides
the missing elements: a huge industrial-scale evidentiary sample
validated by a large group of professional translators across multiple
languages based on professional translation work done in real-world
production scenarios.</span></strong></div><div><strong><br /></strong></div><div>The historical difficulty in providing acceptable proof does not mean
that progress is not being made, but it is helpful to place these claims
in proper context and perspective to better understand what the
implications are for the professional and enterprise use of MT
technology. </div><div><br /></div><div><strong><span style="color: #ffa400;">The history of MT (machine translation) is unfortunately filled with empty promises</span></strong></div><div><strong><br /></strong></div><div>MT (<em>human language translation</em>) is considered among the most
difficult theoretical problems in AI, and thus we should not be
surprised that it is a challenge that has not yielded completely to the
continuing research efforts of MT technology experts over the decades.
Also, many experts have said that MT is a difficult enough challenge (<em>AI-complete:
because it requires a deep contextual understanding of the data, and
the ability to make accurate predictions based on that data</em>) that it is a good proxy for AGI (<em>Artificial
general intelligence is the ability of a machine process/agent to
understand or learn any intellectual task that a human being can</em>) and thus progress with MT can also mean that we are that much closer to reaching AGI.</div><div><br /></div><div><h1 style="text-align: left;">The Historical Lack of Compelling Evidence</h1></div><div><p>MT researchers are forced to draw conclusions on research progress
being made based on relatively small samples of non-representative data (<em>from the professional translation industry perspective</em>) that are evaluated by low-cost human "translators". The <a href="https://kv-emptypages.blogspot.com/2016/09/the-google-neural-machine-translation.html">Google Translate claims in 2016</a>
are an example of a major technology developer making
"human-equivalence" claims based on limited data that was possible
within the scope of the technology development process typical at the
time. </p><p>Namely, here are 200 sentences that amateur translators say
are as good as human translation, thus we claim we have reached human
equivalence with our MT.</p><p>Thus, while Google did indeed make
substantial progress with its MT technology, the evidence it provided to
make the extravagant claim lacked professional validation, was limited
only to a small set of news domain sentence samples, and was not
representative of the diverse and broad scope of typical professional
translation work which tends to be much more demanding and varied.</p><p>The
problem from the perspective of the professional industry with these
historical as-good-as-humans claims can be summarized as follows:</p><ol><li><strong>Very small samples of non-representative data</strong>:
Human equivalence is claimed on the basis of evaluations of a few news
domain segments where non-professional translators were unable to
discern meaningful differences between MT and human translations. The
samples used to draw these conclusions were typically based on no more
than a few hundred sentences.</li><li><strong>Automated quality metrics like BLEU were used to make performance claims: </strong>The
small samples of human evaluation were generally supported by larger (a
thousand or so) sentences where the quality was assessed by an
automatic reference-based score. There are many problems with these
automated quality scores as <a href="https://blog.modernmt.com/understanding-mt-quality-bleu-scores/">described here</a>,
and we now know that they miss much of the nuance and variation that is
typical in human language, resulting in erroneous conclusions, and at
best they are very rough approximations of competent human assessments. <strong>COMET
and other metrics are slightly better quality approximation scores but
still fall short of competent human assessments which are still the
"gold standard" in assessing translation output quality.</strong> The
assessments of barely bilingual translators found in Mechanical Turk
settings and often used by MT researchers are likely to be quite
different from expert professional translators whose reputations are
defined by their work product. <strong>Competent human assessments
("gold standard") are often at odds and different from the segments
suggested as the best-scoring ones based on metrics like COMET or
hLepor.</strong></li><li><strong>Overreaching extrapolations:</strong>
The limited evidence from these experiments was marketed as
“human-equivalence” by Google and others, and invariably resulted in
disappointing professional translators and enterprise users who quickly
witnessed the poor performance of these systems when they strayed away
from news domain content. <strong>Though these claims were not
deliberately deceptive, they were made to document progress from a
perspective that was much narrower than the scope and coverage typical
of professional translation work. </strong>There has never been a claim
of improved MT quality performance based on the huge scale (across 2
billion segments) presented by Translated SRL.</li></ol><div><br /></div><div><h3 style="text-align: left;"><span style="font-size: medium;">Translated SRL Finally Provides Compelling Evidence</span></h3></div></div><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi_hr8sDS0YOphecnl85AOsAmib55rPF9uVesX0N919AzrZ1qJru_bNNjLHkwFwM1rDyj5D_eAM5dM-C3UcIHFm-v3vib7UjFctMexbdDAUHhnbWWt-Z4PM5zX5PSKUnBNWgxIm5ndKkx3fUkwY7tRRfxZuG8wF8ZRGOwwRnfgC7V8QSy7hleaM2Oz7Wg/s1980/data-showing-speed-to-singularity.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="1040" data-original-width="1980" height="210" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEi_hr8sDS0YOphecnl85AOsAmib55rPF9uVesX0N919AzrZ1qJru_bNNjLHkwFwM1rDyj5D_eAM5dM-C3UcIHFm-v3vib7UjFctMexbdDAUHhnbWWt-Z4PM5zX5PSKUnBNWgxIm5ndKkx3fUkwY7tRRfxZuG8wF8ZRGOwwRnfgC7V8QSy7hleaM2Oz7Wg/w400-h210/data-showing-speed-to-singularity.jpg" width="400" /></a></div><div class="separator" style="clear: both; text-align: center;"><br /></div><div class="separator" style="clear: both; text-align: left;"><p>The measurement used to describe ongoing progress with MT is Time To
Edit (TTE). This is a measurement made during routine production
translation work and represents the time required by the world’s
highest-performing professional translators to check and correct
MT-suggested translations. </p><p>Translated makes extensive use of MT
in their production translation work and has found that TTE is a much
better proxy for MT quality than measures like Edit Distance, COMET, or
BLEU. They have found that rather than using these automated score-based
metrics, <strong>it is more accurate and reliable to use a measurement
of the actual cognitive effort extended by professional translators
during the performance of production work. </strong></p><p><strong>Consistent
scoring and quality measurement are challenging in the production
setting because this is greatly influenced by varying content types,
translator competence, and changing turnaround time expectations</strong>.
A decade of careful monitoring of the production use of MT has yielded
the data shown above. Translators were not coerced to use MT and it was
only used when it was useful. </p><p>The data are compelling because of the following reasons:</p><ul><li>The <strong>sheer scale of the measurements across actual production work</strong>
is described in the link above. The chart focuses on measurements
across 2 billion edits where long-term performance data was available.
</li><li>The chart represents what has been <strong>observed over seven years, across multiple languages, measuring the experience of professional translators </strong>making about 2 billion segment edits under real-life production deadlines and delivery expectations.</li><li><strong>Over 130,000 carefully selected professional translators contributed to the summary measurements shown on the chart.</strong></li><li><strong>The
segments used in the measurements are all no TM match segments as this
represents the primary challenge in the professional use of MT.</strong></li><li>The broader ModernMT experience also shows that <strong>highly optimized MT systems for large enterprise clients are already outperforming the sample shown </strong>above which represents the most difficult use case of no TM match.</li><li><strong>A
very definite linear trend shows that if the rate of progress continues
as shown, it MAY be possible to produce MT segments that are as good as
those produced by professional translators within this decade.</strong>
This is the point of singularity at which the time top professionals
spend checking a translation produced by the MT is not different from
the time spent checking a translation produced by their professional
colleagues which may or may not require editing.</li></ul><div><h3 id="it-is-important-to-understand-that-the-productivity-progress-shown-here-is-highly-dependent-on-the-superior-architecture-of-the-underlying-modernmt-technology-which-learns-dynamically-and-continuously-and-improves-on-a-daily-basis-based-on-ongoing-corrective-feedback-from-expert-translators-modernmt-output-has-thus-continued-to-steadily-improve-over-time-it-is-also-highly-dependent-on-the-operational-efficiency-of-the-overall-translation-production-infrastructure-at-translated-srl"><span style="color: #2b00fe;">It
is important to understand that the productivity progress shown here is
highly dependent on the superior architecture of the underlying
ModernMT technology which learns dynamically, and continuously, and
improves on a daily basis based on ongoing corrective feedback from
expert translators. ModernMT output has thus continued to steadily
improve over time. It is also highly dependent on the operational
efficiency of the overall translation production infrastructure at
Translated SRL.</span></h3></div><div><p>The <strong>virtuous data improvement cycle that is created by engaged expert translators </strong>providing regular corrective feedback provides the right kind of data to drive ongoing improvements in MT output quality. <strong>This
improvement rate is not easily replicated by public MT engines and
periodic bulk customization processes that are typical in the industry.</strong></p><p>The
corrective input is professional peer revision during the translation
process - and this expert human input "has control," and guides the
ongoing improvement of the MT, not vice versa. <strong>While overall
data, computing, and algorithms are critical technological foundations
to ongoing success, expert feedback has a substantial impact on the
performance improvements seen in MT output quality.</strong> </p><p>The
final quality of translations delivered to customers is measured by a
metric called EPT (Errors per thousand words) which in most cases is 5
or even as low as 2 when two rounds of human review are used. <strong>The
EPT rating provides a customer-validated objective measure of quality
that is respected in the industry, even for purely human translation
product when no MT is used. </strong></p><p>There is a strong,
symbiotic, and mutually beneficial relationship between the MT and the
engaged expert translators who work with the technology. The process is
quite different from typical clean-up-the-mess PEMT projects with
customized static models where the feedback loop is virtually
non-existent, and where the MT systems barely improve even with large
volumes of post-edited data.</p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjlFALvrGzVgz8CxuPGMQig-winKbEwXh3TpqGyt4S969BDlmAPZj55FasFyfwmi1WpfZuaZyWAMUo7GaF0-9bxl06BbOlb6f3cEXaT0aQob1o-tXYPE2YTojaJGnmpE9pPLSR4_6_Rd7LRRgbueEBAWoU-kvjeMHOFKz10ZhdZxRkaafhH3-Lzif6yog/s910/The-March-Towards-AI-Singularity-and-Why-It-Matters-33-1.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="470" data-original-width="910" height="206" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjlFALvrGzVgz8CxuPGMQig-winKbEwXh3TpqGyt4S969BDlmAPZj55FasFyfwmi1WpfZuaZyWAMUo7GaF0-9bxl06BbOlb6f3cEXaT0aQob1o-tXYPE2YTojaJGnmpE9pPLSR4_6_Rd7LRRgbueEBAWoU-kvjeMHOFKz10ZhdZxRkaafhH3-Lzif6yog/w400-h206/The-March-Towards-AI-Singularity-and-Why-It-Matters-33-1.png" width="400" /></a></div><span style="font-size: x-small;"><div style="text-align: center;">Responsive, Continuously Improving MT Drives Engagement from Expert
Translators </div></span></div><div style="text-align: center;"><span style="font-size: x-small;">Who See Immediate Benefit During the Work Process</span></div><div style="text-align: center;"><br /></div><div><h3 style="text-align: left;">The Problem with Industry Standard Automated Metrics for MT Quality Assessment</h3><div><p>It has become fashionable in the last few years to use automated MT
quality measurement scores like BLEU, Edit Distance, hLepor, and COMET
as a basis to select the “best” MT systems for production work. And some
companies use different MT systems for different languages in an
attempt to maximize MT contributions to production translation needs. <strong>These
scores are all useful for MT system developers to tune and improve MT
systems, however, globalization managers who use this approach may
overlook some rather obvious shortcomings of this approach for MT
selection purposes. </strong></p><p>Here is a summary listing of the shortcomings of this best-MT-based-on-scores approach:</p><ol><li>These scores are typically based on <strong>measurements of static systems</strong>.
The score is ONLY meaningful on a certain day with a certain test set
and actual MT performance may be quite different from what the static
score might suggest. <strong>The score is a measurement of a historical point and is generally not a reliable predictor of future performance.</strong></li><li>Most enterprises need to adapt the system to their specific content/domain and thus <strong>the ability of a system to rapidly, easily, and efficiently adapt to enterprise content is usually much more important</strong> than any score on a given day.</li><li><strong><strong>These
scores do not and can not factor in the daily performance improvements
that would be typical of an adaptive, dynamically, and continuously
improving system like ModernMT</strong>,<strong> which would most likely score higher every day it was actively used and </strong>provided with<strong> corrective feedback. Thus, they are of very limited value with such a system. </strong></strong></li><li>These
scores can vary significantly with the test set that is used to
generate the score and scores can vary significantly as test sets are
changed. <strong>The cost of generating robust and relevant test sets often compromises the testing process as the test process can be gamed.</strong></li><li>Most of these scores are only<strong> based on small test sets with only 500 or so sentences</strong>
and the actual experience in production use on customer data could vary
dramatically from what a score based on a tiny sample might suggest.</li><li>Averaged over many millions of segments, <strong>TTE
gives an accurate quality estimate with low variance and is a more
reliable indicator of quality issues in production MT use.</strong>
Machine translation researchers have had to rely on automated
score-based quality estimates such as the edit distance, or
reference-based quality scores like COMET and BLEU to get quick and
dirty MT quality estimates because they have not yet had the opportunity
to work with such large (millions of sentences) quantities of data
collected and monitored in production settings. </li><li>As enterprise
use of MT evolves the needs and the expected capabilities of the system
will also change and thus static scores become less and less relevant to
the demands of changing needs.</li><li>Also, such a score does not
incorporate the importance of overall business requirements in an
enterprise use scenario where other workflow-related, integration, and
process-related factors may actually be much more important than small
differences in scores.</li><li>Leading-edge research presented at EMNLP 2022 and similar conferences provide evidence that <strong>COMET-optimized
system rankings frequently do not match what “gold-standard” human
assessments would suggest as optimal. Properly done human assessments
are always more reliable in almost every area of NLP.</strong> The TTE
measurements described above inherently allow us to capture human
cognition impact and quality assessment at a massive scale in a way that
no score or QE metric can today.</li><li>Different MT systems respond
to adaptation and customization efforts in different ways. The benefit
or lack thereof from these efforts can vary greatly from system to
system especially when a system is designed to primarily be a generic
system. <strong>Adaptive MT systems like ModernMT are designed from the
outset to be tuned easily and quickly with small amounts of data to fit a
wide range of unique enterprise use cases</strong>. ModernMT is almost
never used without some adaptation effort, unlike generic public MT
systems like Google MT which are primarily used in a default generic
mode. </li></ol></div><p><br /></p><p><strong><span style="color: #2b00fe;"><strong>A “single point quality score” based on publicly sourced
sentences is simply not representative of the dynamically changing,
customized, and modified potential of an active and evolving enterprise </strong>adaptive <strong>MT system</strong> that is designed to be continuously adapted to unique customer use case requirements.</span></strong></p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkC5A73RJ3dy33gkMp2U_4FBqG7Gu1EynHgx0N1Yhh75uMcL3LPTDwYsT1ZO9Pl7lPA3dsx0kPbdif3dwB3GGdEiwgn560ytXXYxASq73njnSWGEx9bk2HbTc7h3OBoGs-KfPRt88HpS0b5Zw5e1lIiHbO-dpmKqcV2D2RBIIw8cdNWoX_mUE3bsZF8Q/s1249/The-March-Towards-AI-Singularity-and-Why-It-Matters-36.png" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="470" data-original-width="1249" height="150" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjkC5A73RJ3dy33gkMp2U_4FBqG7Gu1EynHgx0N1Yhh75uMcL3LPTDwYsT1ZO9Pl7lPA3dsx0kPbdif3dwB3GGdEiwgn560ytXXYxASq73njnSWGEx9bk2HbTc7h3OBoGs-KfPRt88HpS0b5Zw5e1lIiHbO-dpmKqcV2D2RBIIw8cdNWoX_mUE3bsZF8Q/w400-h150/The-March-Towards-AI-Singularity-and-Why-It-Matters-36.png" width="400" /></a></div><strong><p><strong><br /></strong></p>When it is necessary to compare two MT systems in a buyer
selection & evaluation process, double-blind A/B human evaluations
on actual client content would probably produce the most accurate and
useful results that are also better understood by the executive and
purchasing management.</strong><p></p><p><strong>Additionally, MT systems are not static: the models are
constantly being improved and evolving, and what was true yesterday in
quality comparisons may not be true tomorrow.</strong> For these reasons, understanding how the <strong>data, algorithms, and human processes</strong> around the technology interact is usually more important than any static score-based comparison snapshot. A more detailed <a href="https://blog.modernmt.com/understanding-mt-quality/">discussion of the overall MT system comparison issues is provided here</a>.</p><p>Conducting
accurate and consistent comparative testing of MT systems is difficult
with either automated metrics or human assessments. We are aware that
the industry struggles in its communications about translation quality
with buyers. Both are easy to do badly and difficult to do well.
However, in most cases, properly done human A/B tests will yield much
more accurate results than automated metrics.</p><p> <b><span style="color: #2b00fe;">Questions to ask when looking at automated metrics: </span></b></p><p></p><ul style="text-align: left;"><li><b><span style="color: #2b00fe;">What specific data
was used to calculate the score? </span></b></li><li><b><span style="color: #2b00fe;">How similar or different is it from my
data? </span></b></li><li><b><span style="color: #2b00fe;">Can I see the data that was used? </span></b></li><li><b><span style="color: #2b00fe;">How easy or difficult is it to
adapt this MT system to my specific linguistic style and preferences? </span></b></li><li><b><span style="color: #2b00fe;"> How much effort is needed to teach this MT system to use my preferred
style and language? </span></b></li><li><b><span style="color: #2b00fe;">Will I need ML experts to do this or can my
translators drive this? </span></b></li><li><b><span style="color: #2b00fe;">Do small score differences really mean anything? </span></b></li><li><b><span style="color: #2b00fe;"> What happens to these scores if I make changes to the test set? </span></b></li><li><b><span style="color: #2b00fe;">How
quickly will this MT system improve as my translators provide daily
corrections? </span></b></li><li><b><span style="color: #2b00fe;">Do my translators accept these score-based rankings if I
show them the output from 3 different systems? </span></b></li><li><b><span style="color: #2b00fe;">Do my translators like
working with this MT system? </span></b></li><li><b><span style="color: #2b00fe;">Will I be forced to use less qualified
translators if I use this MT system as the best translators will prefer
to decline?</span></b></li></ul><p></p><h2 id="the-implications-of-continuously-improving-mt"><br /></h2><h3 style="text-align: left;">The Implications of Continuously Improving MT</h3><p>Modern
commerce is increasingly done with the support of online marketplaces
and the importance of providing increasingly larger volumes of relevant
content digitally to customers has become an important requirement for
success. </p><p>As the volumes of content grow, the need for more
translation also grows substantially. Gone are the days when it was
enough for a global enterprise to provide limited, relatively static
localization content.</p><p><strong>Delivering superior customer
experience (CX) requires much more content to be made available to
global customers who have the same informational requirements as
customers in the HQ country do. A deep and comprehensive digital
presence that provides a broad range of relevant content to a buyer and
global customer may be, even more, important to be successful in
international markets.</strong></p><p>The modern era requires huge
volumes of content to support the increasingly digital buyer and
customer journey. Thus, the need for high-quality, easily adapted
machine translation grows in importance for any enterprise with global
ambitions. </p><p>The success and relentless progress of the ModernMT
technology described here make it an ideal foundation for building a
rapidly growing base of multilingual content without compromising too
much on the quality of translations delivered to delight global
customers. <strong>This is critical technology needed to allow an
enterprise to go multilingual at scale. This means that it is possible
to translate billions of words a month at relatively high quality.</strong></p><p></p><div class="separator" style="clear: both; text-align: center;"><a href="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5e63xB0ykqXYAj8ZxzqPTpGCZAKEz8lSFqDmfW9TPLvnKXFBx3LkAceOI3o2LPxCGLEIicH4jUy8y9ThY8kBrEKTU5spGmTiIaIUcrc_dNy7xjNWMxPbOvpfv-5GrCLtSKKnja7GmU384h4BMD85D31dyl7t2f0jqQsIbULrXJ06WJHPJCplmUMEL9A/s986/CX%20Age%203.jpg" imageanchor="1" style="margin-left: 1em; margin-right: 1em;"><img border="0" data-original-height="569" data-original-width="986" height="231" src="https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEh5e63xB0ykqXYAj8ZxzqPTpGCZAKEz8lSFqDmfW9TPLvnKXFBx3LkAceOI3o2LPxCGLEIicH4jUy8y9ThY8kBrEKTU5spGmTiIaIUcrc_dNy7xjNWMxPbOvpfv-5GrCLtSKKnja7GmU384h4BMD85D31dyl7t2f0jqQsIbULrXJ06WJHPJCplmUMEL9A/w400-h231/CX%20Age%203.jpg" width="400" /></a></div><br /><strong><br /></strong><p></p><p>The availability of adaptive, highly responsive MT also enables new kinds of knowledge sharing to take place.</p><p>A case in point: <a href="https://translated.com/breaking-language-barriers-in-healthcare">Unicamullus Medical University in Rome experimented with using ModernMT</a>
to translate their medical journal into several new languages and test
acceptance and usability. They were surprised to find that the MT
quality was much better than expected. The success of the initial tests
was promising enough to encourage it to expand the experiment and make
valuable medical journal content available in 28 languages.<strong> </strong></p><p><strong>The
project also allows human corrective feedback to be added to the
publishing cycle when needed or requested. This machine-first and
human-optimized approach is likely to become an increasingly important
approach to large-scale translation needs when intelligent adaptive MT
is the foundation. </strong></p><p>It is quite likely that we will see
possibly 1000X or more growth, in the content volume that is translated
in the years to come, but also that we see a growing use of adaptive and
responsive MT systems like ModernMT which are deeply integrated with
active system-improving human feedback loops that can enable and drive
this massive multilingual expansion. </p><p><strong>There is increasing
evidence that the best-performing AI systems across many areas in NLP
have a well-engineered and tightly integrated human-in-the-loop to
ensure optimal results in production use scenarios. The Translated SRL
experience with ModernMT is proof of what can happen when this is done
well.</strong></p><p>We should expect to see many more global companies
translating hundreds of millions of words a month in the near future to
serve their global customers. A future that will increasingly be
machine-first and human-optimized.</p><p>The following interview with
Translated CEO, Marco Trombetti, provides additional insight into the
progress that we have witnessed with MT over a decade of careful
observation and measurement. <strong>The interview highlights the many
steps taken to ensure that all the measurements are useful KPIs in a
professional translation services setting which has been and will
continue to be the most demanding arena of performance for MT technology</strong>.
Marco also points out that ModernMT and new Generative AI like ChatGPT
are made of the same DNA, and that MT research has provided the critical
technological building blocks used to make these LLMs (Large Language
Models like ChatGPT) possible.</p><p> </p><div class="separator" style="clear: both; text-align: center;"><iframe allowfullscreen="" class="BLOG_video_class" height="266" src="https://www.youtube.com/embed/4kSPH-748ac" width="320" youtube-src-id="4kSPH-748ac"></iframe></div><br /><p></p></div></div><br /><div><br /></div>Kirti Vasheehttp://www.blogger.com/profile/16795076802721564830noreply@blogger.com0