A certain type of ambition doesn’t make a big announcement. For the majority of his career, Christopher Manning has worked on a project that most people outside of computer science wouldn’t even consider challenging: teaching machines to comprehend the true meaning of words. Not merely acknowledge them. Not just spit back results and match syllables in a pattern. When you’re irritated, unsure, or slightly exaggerating for effect, the person seated across from you at a coffee shop will understand.
A few years ago, Manning, a linguistics and computer science professor at Stanford University, was named the Thomas M. Siebel Professor in Machine Learning. It’s the type of honorific that sounds great at a faculty reception but falls short of accurately describing the nature of the work. To put it simply, Manning works to bridge the gap between human language use and machine language processing. It turns out that the gap is huge.
The early days of natural language processing were almost endearing in their adherence to rules. Researchers thought that a computer would eventually follow along if you could write out the grammar exactly enough—subject, verb, object, and clause after clause. It was ineffective. It turns out that language functions very differently from a grammar textbook. Instead of saying “good morning, how are you this morning,” people say “hey, how’s it going” because the rules were never really driving the bus in the first place, not because they have abandoned them.
Data was what made a difference. When the internet suddenly made large amounts of human writing accessible in the 1990s, scholars like Manning began developing probabilistic models, which are essentially systems that could predict which words were likely to follow which other words and what those patterns revealed about meaning. In hindsight, this conceptual change seems clear, but at the time, it wasn’t at all. Language became something to be understood statistically, much like you might understand a crowd by observing its movements, rather than something to be decoded like a cipher.
Machine translation, or the challenge of getting a computer to translate between two human languages without losing the original meaning, has been Manning’s specific area of interest. There may not be a more difficult computational linguistics problem. Every language has its own untranslatable rhythms, internal logic, and cultural shorthand. Around 2005, Google Translate introduced its first statistical system, which was helpful in the same way as a rough sketch: you could get the general idea, but you wouldn’t trust it with anything crucial.
Deep learning and what scientists refer to as neural machine translation—developed at institutions like Stanford and the University of Montreal—provided the breakthrough. The industry was drastically altered overnight by the stark difference in output quality. Baidu, Microsoft, and Google all switched. Manning is open about the fact that the issue is still unresolved and most likely won’t be for decades. That admission from someone who has dedicated his career to it has an almost comforting quality.
Beneath all of this is a question that Manning frequently poses: how can a computer comprehend the world beyond the words that describe it? His response has been to rely on the written record of human experience, which consists of billions of documents, articles, and conversations that encode people’s knowledge and ways of thinking. The model’s depiction of reality is enhanced by each sentence that discusses a hurricane, a scientific finding, or a political dispute. According to his description, the dream is a system that can create its own knowledge structures just by reading, much like an inquisitive person might.
It’s difficult to ignore what Manning and his colleagues were debating in Stanford conference rooms years before any of the current wave of AI language tools became widely used. All of it stems, in one way or another, from the fundamental question he began posing decades ago. This includes voice assistants that only partially comprehend your requests, translation apps that generally get the tone right, and chatbots that sometimes say something surprisingly appropriate. What is the true meaning of this word? How could a machine possibly know?
London Bilingualism's content on health, medicine, and weight loss is solely meant for general educational and informational purposes. This website does not offer any diagnosis, treatment recommendations, or medical advice.
We consistently compile and disseminate the most recent information, findings, and advancements from the medical, health, and weight loss sectors. When content contains opinions, commentary, or viewpoints from professionals, industry leaders, or other people, it is published exactly as it is and reflects those people's opinions rather than London Bilingualism's editorial stance.
We strongly advise all readers to consult a qualified medical professional before acting on any medical, health, dietary, or pharmaceutical information found on this website. Since every person's health situation is different, only a qualified healthcare provider who is familiar with your medical history can offer you advice that is suitable for you.
In a similar vein, any legal, regulatory, or compliance-related information found on this platform is provided solely for informational purposes and should not be used without first obtaining independent legal counsel from a licensed attorney.
You understand and agree that London Bilingualism, its editors, contributors, and affiliated parties are not responsible for any decisions made using the information on this website.
