The Nobel Prize in Physics Goes to Two Researchers for the Foundations of Machine Learning

John Hopfield and Geoffrey Hinton were honored for work in the 1980s that taught computers to learn from data using networks modeled loosely on the brain. Hinton, who quit Google last year to warn about AI, said he was 'flabbergasted.'

The Royal Swedish Academy of Sciences awarded this year's Nobel Prize in Physics to John Hopfield of Princeton and Geoffrey Hinton of the University of Toronto, "for foundational discoveries and inventions that enable machine learning with artificial neural networks." It is the first Nobel given for the science underneath modern artificial intelligence, and it went to two men who did their key work forty years ago, when almost nobody thought it would lead anywhere.

Hopfield, in 1982, built a network that could store patterns and retrieve them from partial clues, the way a person recalls a song from a few notes. Hinton, a few years later, extended it into a network that could learn to recognize features in data on its own. Those ideas, refined over decades and eventually run on thousands of computer chips, are the ancestors of every program described on this site.

Hinton, who is 76, said he was "flabbergasted" when the call came at a hotel in California. He is also the man who resigned from Google eighteen months ago to say publicly that he thought the technology might get out of human control. He said so again today: he compared the coming changes to the Industrial Revolution, said the systems would eventually exceed people in intellectual ability, and added, "We have no experience of what it's like to have things smarter than us."

The choice of the physics prize raised some eyebrows — the work is closer to computer science — and the academy's answer was that the methods came from physics and now serve it. Tomorrow's chemistry prize may make a related point.