Demis Hassabis won the Nobel Prize in Chemistry, and although his work at DeepMind is much more recent, he has previously worked in the gaming industry!
Hassabis co-founded DeepMind, but previously worked as a programmer at Bullfrog and Lionhead. Now, along with professors John Jumper and David Baker, he has won the Nobel Prize in Chemistry. Hassabis and Jumper were awarded the prize for their “complete revolution” in protein structure prediction using the AlphaFold2 artificial intelligence tool.
AlphaFold was first made available to researchers in 2018. The second version will be released in 2020, and AlphaFold3 was announced in May this year. Developed by DeepMind, the artificial intelligence for predicting protein structures is credited to David Baker and his researchers. It has now been used to predict the structure of almost all known proteins, something unimaginable less than a decade ago. Half of the Nobel Prize goes to Professor Baker for “computational protein design”, while the other half is shared by Professor Hassabis and Professor Jumper of DeepMind for “protein structure prediction”. The winners will receive a share of SEK 11 million ($1.05 million).
“One of the discoveries being recognized this year involves the construction of spectacular proteins. The other fulfills a 50-year-old dream: predicting protein structures from their amino acid sequences. Both discoveries open up enormous possibilities,” says Heiner Linke, Chairman of the Nobel Committee for Chemistry. Nobel’s press release defines AlphaFold2 as follows: “AlphaFold2 has been used to predict the structure of virtually all of the 200 million proteins that researchers have identified. Since its breakthrough, AlphaFold2 has been used by more than two million people from 190 countries. Among a myriad of scientific applications, researchers can now better understand antibiotic resistance and create images of enzymes that can break down plastic. Without proteins, life could not exist. The fact that we can now predict protein structures and design our own proteins is the greatest benefit to humanity.”
Hassabis worked at Bullfrog before going to university, and then at Lionhead after graduating. He was the lead AI programmer on Black And White, which was released in 2001, but by then he had moved on to Elixir Studios (which he founded), where he served as executive director of Republic: The Revolution and Evil Genius. From there, he turned his attention to science, earning a PhD, visiting several universities, and co-founding DeepMind in 2010. The goal from the start was to create a general-purpose artificial intelligence that could be used for anything, but it started with video games in the ’70s and ’80s. In the early days, DeepMind’s AI trained on games like Breakout and Pong, learning the rules to master the game, before the company focused on more complex games like Go or even Starcraft 2.
The company was acquired by Google in 2014. Around 2016, with the video game training behind it, the company focused on protein folding and AlphaFold, software that has made impressive advances in protein folding almost from the start. So congratulations to Hassabis!
Leave a Reply