Demis Hassabis from chess genius to Nobel Prize-winning AI pioneer

Long before Demis Hassabis won the Nobel Prize for pioneering artificial intelligence technology, he was a master of board games.

The London-born son of a Greek Cypriot father and a Singaporean mother started playing chess at the age of four and reached the rank of master at 13.

“Playing chess from an early age and trying to improve my own thinking processes is what got me interested in AI in the first place,” said the 48-year-old, who won the Nobel Prize in Chemistry. told journalists after sharing it with two other scientists. Wednesday.

This is the second Nobel Prize win related to artificial intelligence (AI) in recent days, with Hassabis following Tuesday’s chemistry prize winners, who said the technology they championed could be used to “do harm. ” warned that there was a possibility.

But rather than a dire and bleak warning about the apocalypse of AI, the CEO of Google’s DeepMind lab described himself as a “cautious optimist.”

“I’ve been working on this my whole life because I believe this will be the most beneficial technology for humanity. But with something this powerful and innovative, there are risks,” he said.

try your hand at video games

Hassabis left high school in north London at age 16, took a gap year to work on video games, and co-designed 1994’s Theme Park.

In his 20s, Hassabis won Pentamind, a London event that combines the results of bridge, chess, Go, Mastermind and Scrabble, five times.

“I actually encourage kids to play games, but not just play, most importantly try to make games,” Hassabis said.

He then studied neuroscience at University College London, hoping to learn more about the human brain with the aim of improving early AI.

In 2007, Science magazine named his research among the top 10 breakthroughs of the year.

He co-founded a company called DeepMind in 2010 that went on to use artificial neural networks (loosely based on the human brain and powering AI) to beat humans at board and video games. focused on.

Google acquired the company four years later.

In 2016, DeepMind became known around the world after its AI-driven computer program AlphaZero defeated the world’s top players of Go, an ancient Chinese board game.

A year later, AlphaZero defeated the world champion chess program Stockfish, showing that it was no longer a one-game wonder. I’ve also conquered some retro video games.

The point was not to have fun or win the game, but to extend the capabilities of the AI.

“This type of learning technology is what has fueled the modern AI renaissance,” Hassabis said.

protein power

Hassabis then turned the power he had built up into proteins.

These are the building blocks of life, taking information from the DNA blueprint and turning cells into specific things, such as brain cells or muscle cells, or almost anything else.

By the late 1960s, chemists knew that the sequence of the 20 amino acids that make up a protein should be able to predict the three-dimensional structure in which the protein twists and folds.

But for half a century, no one could accurately predict these 3D structures. There was even a bi-annual competition called the “Protein Olympics” for chemists to test their skills.

In 2018, Hassabis and his AlphaFold entered the competition.

Previous predictions had an accuracy rate of at most 40%. AlphaFold exceeded nearly 60%.

Two years later, business performance was so strong that it seemed that the 50-year-old problem had been resolved.

John Jumper of DeepMind, who shared Wednesday’s Nobel Prize with American biochemist David Baker, said about 30,000 scientific papers now cite the alpha fold.

“AlphaFold is already being used by more than 2 million researchers to advance critical research from enzyme design to drug discovery,” said Hassabis.

© 2024 AFP