Wednesday, December 29, 2021

The Human Space Beyond Language

Much of what I write about in this blog is about language technology and machine translation. The primary focus is on the technology and AI initiatives related to human language translation. This focus will remain so, but I recently came upon something that I felt was worth mentioning, especially in this holiday season, where many of us review, consider and express gratitude for the plenitude in our lives.

Language is a quintessentially human experience where we share, discover, learn, and express the many different facets of our lives through this medium we call language. This is probably why computers are unlikely to ever unravel it fully, there is too much amorphous but critical context about life, living, learning, and the world, around most words to easily capture with training data and give to a computer to learn.  

While many of us surmise that language is only about words and about how words can be strung together to share, express, understand the world around us, in most cases, there is much that is unspoken or not directly referenced that also needs to be considered to understand any set of words accurately and faithfully. Sometimes the feeling and emotion are enough and words are not needed.

In 2021 Large Language Models (LLMs) were a big deal and GPT-3, in particular, was all over the news as a symbol of breakthrough AI that to some suggests that a sentient machine is close at hand. Until you look more closely and see that much of what is produced by LLMs are crude pattern reflections that are completely devoid of understanding, comprehension, or cognition in any meaningful sense. The initial enthusiasm for GPT-3 has been followed by increasing concern as people have realized how these systems are prone to producing unpredictable obscenity, prejudiced remarks, misinformation, and so forth. The toxicity and bias inherent in these systems will not be easily overcome without strategies that involve more than more data and more compute.

It is very likely that we will see these increasingly larger LLMs go through the same cycles of over-promising and under-delivering that machine translation has gone through for over 70 years now. 

The problem is the same, the words used to train AI alone do not contain everything needed to establish understanding, comprehension, and cognition. And IMO simply training a deep learning algorithm with many more trillions of words will not somehow create understanding and cognition or even common sense.  

The inability for AI "to understand" was clearly shown by Amazon Alexa recently when it told a child to essentially electrocute herself. "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 to this incident. GPT-3 has also advised suicidal humans to kill themselves in experiments conducted elsewhere. 

The machine is not malicious, it simply has no real understanding of the world and life, and lacks common sense. 

The truth is that we are forced to learn to query and instruct Alexa, Siri, and Google Voice so that they can do simple but useful tasks for us. This is "AI" where the human in the loop keeps it basically functional and useful. Expecting any real understanding and comprehension from these systems without many explicit and repeated clarifications is simply not possible in 2021. 

But anyway, I digress, so, I wanted to talk about the areas where humans move beyond language (as in word-based) but yet communicate, share, and express quintessential humanness in the process. 

It is my feeling that entering this space happens most often with music, especially improvised music where there is some uncertainty or unpredictability about the outcome.  Where what happens, happens, often without a plan, but yet still with a clear artistic framework and structural outline. I happen to play the sitar focusing on the Indian Classical music of North India where the "Raga" is the basic blueprint that provides the needed foundations for highly disciplined improvisatory exploration. 

To a great extent what these musicians do is "shape the air" and create something equivalent to sonic sculptures. These sculptures can be pleasing or relaxing in many ways that only humans can understand, and sometimes can be very moving, which means they can trigger emotional release (tears) or establish a deeply emotional presence (left speechless). Often it is not necessary to understand the actual language used in musical performance since there is still a common layer of feeling, emotion, and yearning that all humans can connect and tap into. 

The key difference of this improvisation-heavy approach from a performance of score-based music is that neither the musician nor the audience really knows at the outset how things will turn out. With a score, there is a known and well-defined musical product that both the audience and the musician are aware of and expect. There is more of an elemental structure. However, here too it is possible for an attendee to listen to an unfamiliar language e.g. an operatic aria in Italian, and be deeply moved, even though the audience member speaks no Italian and may have no knowledge of the operatic drama. The connection is made at a feeling and emotional level, not at the word,  language, or idea cognition level.

I came upon this musical performance of a Sufi (a mystical Muslim tradition) song sung by two musical legends on a commercial platform called Coke Studio Pakistan. Musically, this might be considered "fusion" but it is heavily influenced by Indian classical music and it is sung in Urdu (Braj) which is so close to Hindi (Hindustani) that they are virtually the same language, except that Urdu uses much more Persian vocabulary. The original poem was written in Braj Basha an antecedent of both Urdu and modern-day Hindi. 

This particular performance was a rehearsal and was the first time all the musicians were in the same room, but the producers decided it was not possible to improve on this and published it, as is, since it was quite magical and probably impossible to reproduce. 

There are almost 20,000 comments to the video shown below and this comment by Matt Dinopoulos typifies much of the feedback: "It hit my soul on so many levels and just brought me to tears and I don’t even know what they’re saying.The figures of speech “pluck at one’s heartstrings” and “strikes a chord in me” have found a home in our language for just this reason.

The song/poem Chaap Tilak was written by Amir Khusrau the famous poet laureate of the Indian subcontinent considered one of the most versatile poets and prolific prose-writers of the 13th and 14th centuries. He is also considered by some to be a seminal force in the creation of the sitar and the development of the Khyal music form most prevalent in North Indian classical music today. 

This song essentially expresses Khusrau's gratitude, devotion, love, and longing for communion with his Pir (Guru/Spiritual teacher) whose name is Nizam (Nizamuddin Auliya). Sung from the perspective of a young girl awaiting or yearning for her beloved, it is replete with modest yet enchanting symbols, as it celebrates the splendor of losing oneself in love. Both the use of motifs as well as the language itself were deliberate creative choices by Amir Khusrau, to communicate with common people using familiar ideas and aesthetics.

A closer examination of Khusrau's works will reveal that the Beloved in his songs/poems is always the Divine or the Pir. Many poets in India use the perspective of the romantic yearnings of a young maiden for the beloved as an analogy, as the relationship with the Divine is seen as the most intense kind of love. The longing and union they speak of are always about direct contact with the Sacred and so this song should be considered a spiritual lament whose essential intention is to express spiritual love and gratitude. The translations shown in the video are sporadic but still useful.

Chaap Tilak Performance 

At the time of this publishing, the video above had already had 40 million views. Many thanks to the eminent Raymond Doctor for providing this link which provides a full translation, and useful background to better understand the thematic influences and artistic inspiration for this song.

“As long as a spiritual artist respects his craft, peace will prevail. It is wonderful when a singer has a noble cause and spreads the message of love, peace, and brotherhood as presented by our saints, without greed of money or the world. This is the real purpose of qawwali.”             
                                                                                             Rahat Fateh Ali Khan

“Music doesn’t have a language, it’s about the feeling. You have to put a lot of soul into whatever you are making. Music doesn’t work if you’re only doing it for money or professionally. It works only if it’s from the soul. There’s no price to it.”                                                                                                                                                                                                                                    Aima Baig     

"Information is not knowledge. Knowledge is not wisdom. Wisdom is not truth. Truth is not beauty. Beauty is not love. Love is not music. Music is THE BEST.” 

— Frank Zappa 

I played this song for several friends (mostly musicians) who had no familiarity with Indian music and found that several of them were deeply touched, so much so that some had tears streaming and were unable to speak when the song ended. In fact, they were mystified by how strong an emotional reaction they had to this unfamiliar and alien artistic expression. 

This unexpected, often surprising, emotion-heavy reaction is entirely and uniquely human. This kind of listener impact cannot come from musical virtuosity alone which is abundantly present here, the musicians here are also tapping into a deeper sub-strata of feeling and emotion that only exists in and is shared by humans. 

This is the human space beyond language where understanding happens in spite of initial unfamiliarity. There is something in the human psyche that understands and connects to this even if by accidental discovery, and this initial response often leads to a more substantial connection. We could call this learning perhaps, and this is probably how children also gather knowledge about the world. Intensity and connection probably have a more profound impact on students than pedagogy and quite possibly drive intense learning activity in any sphere.

It is interesting that there are many reaction videos on Youtube where music teachers and YT celebrities from around the world share their first reactions to this particular song and other culturally unfamiliar music. Based on the number of these reaction videos, I guess more and more people are exploring and want to share in the larger human musical experience. Some examples:

  • Latina Ceci Dover left speechless (around 4' 50")
  • British rapper reacts in shock and awe (around 6' 05")
  • A deep and informed analysis of the singing technique and mechanics. "If her voice was an animal it would be an eagle."
  • Seda Nur Turkish German was surprised by the emotional connection. (around 3' 15") It also led her to actually visit Pakistan last week, a trip which she is also sharing in her Vlogs.
  • John Cameron left speechless and in tears (around 11' 30')
  • Waleska & Efra discover a new musical paradigm (around 10' 40")
  • Asian dude is blown away (~2' 09"): Oh My Godness he says at 4' 0", and dances to the chorus like a bird (4' 25") Hilarious responses throughout the song.
It is not always necessary to have this level of virtuosity to find this sub-strata space beyond language. Artists often communicate without words, with just a look, or with presence and full attention. I too have participated in impromptu musical conversations, where friends simply gather to converse musically with a very simple outline depending primarily on improvisation and listening to the other. This is an example, that is difficult to reproduce exactly because it captured an instant in time that was unique.

All this to point out that the data we use to build artificial intelligence completely miss these deeper layers of humanness. It is not just about missing a larger subject, common sense, and physical world context, but especially the non-verbal emotional, and feeling layers that also make us human. 

Intelligence is barely understood by humans even after looking at the issue for eons, so how is it even possible to put this into a computer algorithm? What kind of data are we going to use? How do you model emotion and feeling about knowledge, data, information?

Machine learning and computers are likely to radically transform our lives in the coming decade and change our lives in so many ways, but there are some things like the wordless feeling-filled states of human space beyond language, the spiritual sub-strata that underlies consciousness, that I think is simply not within the province of Man to model or perhaps even to understand. It can only be experienced.  


The poem below was copied from Maria Popova's excellent blog about the awe and wonder of being human. Her backgrounder on Rebecca Elson is timely and worth reading, and she has also written about the connection between music and the neurophysiological mechanism of emotion that I recommend. As she points out: "Emotions possess the evanescence of a musical note."


by Rebecca Elson

Returning, like the Earth

To the same point in space,

We go softly to the comfort of destruction,

And consume in flames

A school of fish,

A pair of hens,

A mountain poplar with its moss.

A shiver of sparks sweeps round

The dark shoulder of the Earth,

Frisson of recognition,

Preparation for another voyage,

And our own gentle bubbles

Float curious and mute

Towards the black lake

Boiling with light,

Towards the sharp night

Whistling with sound.

I wish you all a Happy, Healthy, Peaceful, and Prosperous New Year

Thursday, December 16, 2021

The Evolution of Machine Translation Use in the Enterprise

 The modern enterprise with global ambitions is experiencing an evolving view of the value, scope, and need for language translation to enhance and build global business momentum.

The old view has been to translate what is mandatory to participate in a global market, but the new view is increasingly about listening, understanding, sharing, communicating, and engaging the global customer. The new view requires a modern global enterprise to translate millions of words a day.

Recent events, the pandemic, in particular, have forced many private and governmental institutions to become much more focused on expanding their digital presence and profile.

It is now widely understood that providing increasing volumes of content to assist, inform, and help a potential buyer understand the products and services of an enterprise, and enhance the customer’s journey after the purchase are important determinants of sustained market success.

The cost of failure to do so is high. In the last 20 years, over half (52%) of Fortune 500 companies have either gone bankrupt, been acquired, or ceased to exist as a result of the digital and business model disruption. Studies suggest many more companies across various industries will disappear because they fail to understand the strategic value of providing relevant content and establishing a robust digital presence to improve CX.

Innosight Research predicts that as much as 75% of today’s S&P 500 will be replaced by 2027.

Rebecca Ray eloquently describes the impact of producing relevant content for the modern enterprise. She describes how Airbnb and Expedia recognize that they are not just lodging companies, but rather high-tech (multilingual) content companies solving lodging problems.

The [most] significant implications revolve around the recognition given to the business value of multilingual content by a high-tech company such as Airbnb through its financial investment in a cross-functional collaboration initiative to greatly expand language accessibility. They must recognize that, in many cases, their products and services do not function independently of information about them and that the most valuable content and code are often generated by third parties.

As more of the world gets more accustomed to a digital-first buyer journey, companies must adapt to stay relevant. As brands face unmatched logistical and communication challenges in the new millennium, they have focused on more engagement with their customers via digital channels.

Customer experience is no longer just a concept. It’s a business imperative that requires cross-functional collaboration, data, and analytics focused on delivering customer success.

The enterprises with the most successful digital transformation initiatives are seeing growth from improved and innovative digital interactions. This usually means cross-organizational collaboration with a digital strategy-focused team leader invested in optimizing digital channels and improving both customer and employee digital experience. Success is often built on data and analytics and the increasing use of AI to generate predictive analytics to power better, more personalized interactions.

The benefits of providing superior CX are increasingly clear:

  • #1 – By 2021, customer experience will overtake price and product as the key brand differentiator – Walker
  • 86%– of those who received a great customer experience were likely to repurchase from the same company; compared to just 13% of those who received a poor CX – Temkin Group
  • 6x– Between 2010-2015, CX leaders grew 6x faster than CX laggards – Forrester
  • “Customer Experience leaders grow revenue faster than CX laggards, drive higher brand preference, and can charge more for their products.” – Forrester’s Rick Parish.
  • 3x greater return CX leaders outperform CX laggards in terms of stock performance. -- Watermark Consulting
  • Consumers will pay a 16% price premium for a great customer experience. - PwC

Leaders develop “digital agility”, that enables cross-function collaboration focused on mapping and optimizing customer journeys. The key difference according to Altimeter surveys between top-performing companies and average performers is the ability to use data in prescriptive ways. This means harnessing analytics to make or automate decisions that improve processes like delivering a great customer experience, creating a new product, or defining a new strategy.

There is increasing convergence between marketing, sales, and service goals and operations. This convergence is the natural outcome of the increasing digital sophistication of all customer-facing functions.

The Localization Implications

There are significant implications from these larger digital transformation imperatives for traditional localization departments. The needs and scope of language translation for the modern digitally-agile enterprise are significantly greater than ever seen historically.

This is resulting in a changing view of the localization function within the enterprise. Ranging from being seen as a vital partner in global growth strategies to sometimes being relegated to low-value contributors (for those less engaged localization groups) whose only focus is measuring translation quality (badly) and juggling LSP vendors. These latter groups will see customer-facing departments take over large-scale CX-focused translation initiatives and slip into further obscurity.

Airbnb is an example where the localization team is seen as a vital partner in enabling global growth. The Airbnb localization team oversees both typical localization content and user-generated content (UGC), across the organization, which means they oversee billions of words a month being translated across 60+ languages using a combined human plus continuously improving MT translation model. The localization team enables Airbnb to translate customer-related content across the organization at scale. High-value external content is often accorded the same attention as internally produced marketing content.

Digital agility requires that a modern era localization team be the hub of cross-functional collaboration to facilitate and enable all kinds of content to be translated to expedite understanding, communication, and experience of customer-related issues. The modern enterprise will likely need millions of words translated every month to understand, enhance, and improve the global customer experience (GCX). In the B2C scenario, the volumes could even be billions of words per month.

Continuously improving responsive MT is a critical foundation to building better GCX. Increasingly we will see more and more content come directly from carefully optimized MT to the customer without additional human intervention. The linguistic human oversight process and approach will likely change. Upfront human feedback investments will be needed in addition to selective post-editing to make these enterprise MT engines perform optimally on a range of unique and specialized enterprise content. It is possible to post-edit 100K words a month, but it is not possible to do this for a million or a billion source words a month.

We are seeing an emerging translation model (e.g. at Airbnb) that enables an enterprise to build virtuous cycles of continuously improving machine translation working closely together with regular ongoing human feedback.

All focused on cross-functional alignment with the larger enterprise goal of reducing customer friction and increasing customer delight.

Superior CX is built on active, ongoing conversations with the customer to ensure the best possible experience and outcomes. It involves active and continuous listening to the voice of the customer on social platforms to identify problems and success early. It involves multiple levels of communication. Large volumes of multilingual data flows have created a huge and growing need for rapid translation.

Thus, the emerging CX-focused and digitally agile modern enterprise is likely to be one that regularly does the following:

  • Helps to provide a uniform customer data profile across the world to achieve a common understanding of the customer.
  • Translates over 100 million or even billions of words a month to support GCX in various ways. This means that 95%+ has to be done by optimized adaptive MT engines.
  • Has informed human feedback driving continuously learning MT systems. This feedback may only amount to 1% or less of the raw MT that is freely flowing to support understanding, communication, and research of customers across the globe.
  • Helps provide a common view on the increasing convergence between marketing, sales, and service content and operations without any language barriers. This has great value for the internal understanding of international customer needs and again enables speedier, better-coordinated responses to differing international customer needs.
  • Improves global understanding and communication within and without the enterprise.

The State of Machine Translation in the Enterprise

Machine translation output quality has improved dramatically over the past decade, but all stakeholders should understand that MT is not a perfect replacement for competent human translation. Yet, if properly used and optimized, it can rapidly enhance the transparency of all multilingual data flows in the modern enterprise and expedite international business initiatives.

The use of MT in the enterprise is clearly on the rise. The overall content deluge, the need to monitor brand feedback on social media, and a much more rapid digitally-driven global presence are some of the factors that drive this use.

Global customer service and support and eCommerce have been the most active use cases historically. But increasingly we see that that the global enterprise recognizes the need for a personalized and optimized MT-based translation utility that is secure, private, and enterprise content-tuned to make ALL information multilingual and easier to share, access, and understand.

Recent Changes Driving Faster MT Adoption

  • Human evaluations show that neural MT is often indiscernible from human translation for much of the high-value customer service content that is needed for CX improvement.
  • Direct customer feedback on the usefulness of MT content suggests that many customers are willing to accept “imperfect” MT if it means that they get broader content access and faster response during many stages of the customer journey.
  • The pandemic has made the need for an expanded digital presence much more urgent. MT enables the acceleration of a global “digital-first” strategy.
  • The increasing awareness at executive levels of the need for global inclusion and recognition that the fastest-growing markets in the world today are in Africa and SE Asia.
  • The increasing awareness in the enterprise of the need and impact of greater availability.
  • The increasing importance of community content is increasingly recognized business-enhancing customer creator content.
  • An understanding that the next billion potential customers coming to the internet are unlikely to speak English, FIGS, or CJK.

One important point of understanding within a global enterprise is to recognize that different kinds of content need different translation processes. MT is very useful in making high volume, rapidly flowing, short shelf-life content multilingual, but is less suited to high impact executive communications or high liability impact types of content where nuance and semantic accuracy are central.

The mix of human-machine contributions tends to be closely linked to the volume and nuance contained in the source data. The following chart roughly shows the relationship between the content type and the translation approach used. The translation production mode has to be tuned to the needs of target consumers of information and could range from assimilation, dissemination to publishable quality requirements. For example, internal cross-lingual emails can stand a lower average quality translation than customer support content even though both can be voluminous.

What Does an Enterprise Need from an MT Solution?

In considering MT technology options there are several attributes that an enterprise needs to understand in an evaluation. While there is often an over-emphasis and focus on generic, mostly irrelevant MT score-based comparisons, in reality, other factors often matter much more in producing higher ROI and successful outcomes. For example:

Data Security & Privacy: As more CX and confidential product information starts flowing through MT, data security and privacy is our primary concern. CX data may often be considered even more valuable than product information. Many generic public portals present challenges in ensuring the assurance of data privacy. Flexible caching options for different content types are an increasingly important concern.

Rapid Adaptability to Enterprise Content and Use Cases: Enterprise use of MT makes the most sense when the MT performs well on unique enterprise content and terminology. The ability to do this quickly and accurately with minimal startup effort is perhaps the single most valuable aspect of MT technology to an enterprise. There are likely to be many use cases and MT systems like ModernMT, that can leverage existing linguistic assets across multiple use cases with minimal overhead, and system redundancy is much easier to manage and update than a lot of traditional MT systems. Continuous improvement capabilities allow MT systems to evolve and improve over time and thus responsive, highly adaptive MT systems like ModernMT that improve daily are much more likely to produce successful outcomes.

Superior MT Output Quality: While generic MT comparisons by third parties can be useful to understand the potential experience with an MT solution, it is much more important to understand how an MT solution performs on your specific enterprise content. Perhaps even more important is how rapidly and easily an MT system can improve and learn your specific enterprise terminology and linguistic style.

Ease of Integrating Human Corrective Feedback: While some MT systems can produce excellent output occasionally, none can do this all the time.

Human corrective feedback is key to improving MT system performance on an ongoing basis. The ease, speed, and impact with which human feedback is incorporated into driving ongoing MT system performance improvements is a valuable characteristic for an MT system.

ModernMT has been architected to learn rapidly from human corrective feedback and learns continuously and rapidly. The low overhead of adapted models also makes it possible to have hundreds of different models that are easily updated and maintained as improvements can flow from model to model if they have similar content.

Expert Consulting Services: As MT-based translation services become more commonplace in the global enterprise, the reach of rapid translation capabilities will expand and extend through an organization.

It may be necessary to connect different content-containing systems to the MT engine or special linguistic analysis may be needed to speed up the development of systems optimized for new use cases.

MT providers who have this expertise to build a Translation Operating System that can handle a variety of different kinds of data and service types with high efficiency will make the deployment easier, and also increase the probability of success.

ModernMT is a context-aware, incremental, and responsive general-purpose MT technology that is price competitive to the big MT portals (Google, Microsoft, Amazon) and is uniquely optimized for LSPs and global enterprises, and addresses all the criteria specified above.

ModernMT can be kept completely secure and private for those willing to make the hardware investments for an on-premise installation. It is also possible to develop a secure and private cloud instance for those who wish to avoid making hardware investments.

ModernMT overcomes technology barriers that hinder the wider adoption of currently available MT software by enterprise users and language service providers:

  • ModernMT is a ready-to-run application that does not require any initial training phase. It incorporates user-supplied resources immediately without needing laborious, technically overly complex, and tedious upfront model training.
  • ModernMT learns continuously and instantly from user feedback and corrections made to MT output as production work is being done. It produces output that improves by the day and even the hour in active-use scenarios.
  • ModernMT manages context automatically and does not require building multiple different domain-specific and use-case-specific systems.
  • ModernMT has a data collection infrastructure that accelerates the process of filling the data gap between baseline systems and enterprise-specific models with unique terminology and linguistic style.
  • ModernMT is responsive to small volumes of corrective feedback and thus allows straightforward deployment in multiple organizational scenarios.
  • ModernMT’s goal is to deliver the quality of multiple custom engines by adapting to the provided context on the fly. This fluidity makes it much easier to manage on an ongoing basis.
  • ModernMT provides extensive control over the MT cache allowing immediate no-trace deletion to a permanent (annual) cache for highly static public content.
  • ModernMT can scale from millions to billions of words a day with responsive and modifiable performance through the range.
  • ModernMT systems can easily outperform competitive systems once adaptation begins, and active corrective feedback immediately generates quality-improving momentum.

Thus as we head into the post-pandemic era, the need for a broadly capable, private, secure, corporate-domain tuned  "translation engine" will only grow in importance for the digitally agile, CX-savvy global enterprise.

Localization teams will need to lead multi-year investment initiatives that are recognized by executives as essential drivers for their organizations’ global growth and revenue to enable this.  

The shift to language as a feature at the platform level wherein language is designed, delivered, and optimized as a feature of a product and/or service from the beginning is now underway.

As Airbnb CEO, Brian Chesky stated in his recent launch video, “Technology made it possible to work from home, but Airbnb now allows you to work from any home.” Obviously, language independence is an integral part of his platform that now makes this possible. His success will demand that others follow.

This post was originally published here