About Geoffrey Hinton

Today is the first day after the National Day, and it is also a very special day.

During the day, the whole country was discussing stocks. In the afternoon, the news of the Nobel Prize in Physics suddenly emerged, which set off a wave of heat again.

As you can see, the winners of this year's Nobel Prize in Physics are John J. Hopfield of Princeton University in the United States and Geoffrey E. Hinton of the University of Toronto in Canada. Both of them are experts in the field of AI. Especially Hinton, he can be said to be a godfather in the academic world of AI.

Before 2024, Xiaozaojun didn't even know the name Hinton. This year, because I also started to contact and learn AI and wrote some articles about AI, I got to know him.

Hinton is a very awesome person. His life is extremely legendary.

He was born into an academic family (1947), and his undergraduate degree was in psychology (1970). After graduation, he lost his way and worked as a carpenter for a while. Later, after stumbling, he returned to the academic world and studied neural networks under a professor at the University of Edinburgh, and got a doctorate in artificial intelligence (1978). Students

who have read my book "A Brief History of Artificial Intelligence" may still remember that human artificial intelligence started in the 1950s, and the first generation of founders were Turing, McCarthy, Minsky and other bigwigs. However, the route taken by the early artificial intelligence academic community was wrong (symbolism, expert system). What's more, Minsky also denied the correct route of neural network early on.

In the 1980s, experts couldn't get through the old road, so they remembered the road of neural network (machine learning) again.

In 1982, John Hopfield (another bigwig who won the Nobel Prize this time) proposed the Hopfield network model (in fact, it was not his original creation, and the prototype of this model was proposed by other scientists in the early days). In 1986, David Rumelhart, Geoffrey Hinton, Ronald Williams and others jointly published a paper titled "Learning representations by back-propagation errors", proposing the back-propagation algorithm. These two achievements laid an important foundation for neural networks (machine learning), and can be regarded as putting artificial intelligence on the right track.

Hinton was 39 years old in 1986. He had actually been living in the shadow of his family. His father had long been PUAing him, saying that he could not achieve his own achievements (professor at Cambridge University, entomologist). Unfortunately, his father died early, and Hinton had no way to prove it to his father.

The story did not end here.

Twenty years later, in 2006, Hinton broke out again. He published an important paper "Reducing the dimensionality of data with neural networks", proposed Deep Belief Networks (DBNs), brought artificial intelligence into the era of "deep learning", and solved the problem of learning efficiency that had plagued AI for a long time at that time. This initiative can be said to be a timely help to the field of AI. In

2012, Jeffrey Hinton and his students Ilya Sutskever and Alex Krizhevsky participated in the annual large-scale visual recognition challenge held by ImageNet (initiated by Fei-Fei Li), and won a landslide victory using deep learning models, shocking the academic community again.

Later, the three masters and apprentices won a lot of honors and wealth, and the company they founded was acquired by Google at a high price.

In the OpenAI "coup" that attracted global attention at the end of last year, the former chief scientist of OpenAI who tried to fire the company's CEO Sam Altman was Ilya Sutskevi, one of his apprentices.

In short, Hinton is definitely a top boss in the field of AI. His two major contributions have completely rewritten the development direction of AI, and he deserves the award. When I have the chance, I will write a detailed personal biography of Hinton for everyone to see.

Now the only question is actually-

Hinton won the Turing Award (2018) before, why did he win the Nobel Prize in Physics this time? Is this treating AI as physics? The physics circle next door is already in an uproar!