In the chess and gaming world at large, Demis Hassabis has been celebrated many times over the past years for the revolutionary and pioneering efforts of DeepMind in chess, Go, and even shogi. Had there been a Nobel prize or similar in chess, he would have received them already, that is certain. The revolutionary neural networks that are the staple of chess engines of all names can be traced in a straight line from AlphaZero's ground-breaking paper and results.

Demis Hassabis is of course also a well-established and bona fide prodigy in chess himself, and at the bright age of 13 was the second highest rated youth in his category, rated 2300 FIDE, second only to the legendary Judit Polgar.

Lawrence Cooper, Demis Hassabis, Cathy Haslinger and Dharshan Kumaran in around 1986. (source: British Chess News)

Of course, when AlphaGo first came out, wowing the world at large with its match against Go legend Lee Sedol, a lot of behind the scenes talk came about on whether something similar might not be possible in chess. I will unabashedly admit that I was among the doubters, sure that the inherently tactical nature of chess and deep calculations would make it an interesting science project at best, but was unlikely to yield a breakthrough result. No need to state just how wrong I was, though I was hardly alone.

It later became the topic of the brilliant documentary, which is a must-see if you haven't already.

Nevertheless, it was also clear that unlike previous efforts to link chess to AI, such as the unforgettable Deep Blue matches, AlphaZero and Co. were never meant to be the final goals, the summit in some gaming AI Everest. They were stepping stones to help prove that AI could learn and master topics with minimal information. In other words, for as revolutionary as they were to our own confined worlds, these were only meant as proofs of concept.

It was obvious that AI, and more specifically machine learning as is the case, had just awoken and could achieve wonders, but what would those be? DeepMind, driven by its founder and CEO Demis Hassabis, aimed its sights on protein folding. A bit like SETI, protein folding had long been a hugely underdeveloped unicorn project in which the nerd world at large would install a program to allow their home computers to freely analyze data whenever their machines were idle, and then forward the results. It was a noble idea, but it seemed to manifest more a dream than a developing reality. The number of combinations correctly decoded was ridiculously small, but what was the alternative? It was in these conditions that AlphaFold was born.

A great short intro to AlphaFold by DeepMind

Proteins underpin every biological process, in every living thing. Made from long chains of amino acids, each has a unique complex 3D structure. But figuring out just one of these can take several years, and hundreds of thousands of dollars.

You might imagine, correctly, that this will have incredible ramifications in medicine and biology, but it goes far beyond such reductive possibilities, however powerful. Consider that 91% of all plastic ever produced has never been recycled. AlphaFold could help us face up to the challenge of cleaning up our world. 40% of the world’s crops are lost to disease each year. AlphaFold could unlock insights that help keep food on tables. The list goes on.

In 2020, AlphaFold solved this problem, with the ability to predict protein structures in minutes, to a remarkable degree of accuracy. Now owned by Google, one might fear what strings could come attached to the indisputable cost in developing and powering this technology, but there were none. 

The AlphaFold2 paper was groundbreaking and has been among the most cited ever. it is available for free access at Nature magazine.

Not only did the DeepMind teams publish papers sharing the work for all to use and benefit from, but they have hosted a free server with over 200 million 3D protein combinations, almost every one known to science so far, so that researchers around the world might access and benefit from the results. To date some two million researchers from 190 countries around the world access and use that data.

The vast database with over 200 million combinations all modeled is freely available to all. 

Not sitting on their laurels, a newer and more powerful effort led to AlphaFold2, a paper that has since become one of the most cited papers in the history of science.

The awarding of the Nobel prize in Chemistry to Demis Hassabis and John Jumper, a lead researcher on this particular effort, could not be more deserved, and we at ChessBase celebrate their deserved accolades and fanfare.

In a statement released after being informed of the news, Demis Hassabis said:

"Receiving the Nobel Prize is the honour of a lifetime. Thank you to the Royal Swedish Academy of Sciences, to John Jumper and the AlphaFold team, the wider DeepMind and Google teams, and to all my colleagues past and present that made this moment possible. I’ve dedicated my career to advancing AI because of its unparalleled potential to improve the lives of billions of people. AlphaFold has already been used by more than two million researchers to advance critical work, from enzyme design to drug discovery. I hope we'll look back on AlphaFold as the first proof point of AI's incredible potential to accelerate scientific discovery."

Links

• The future is here – AlphaZero learns chess
• AlphaGo vs Lee Sedol: history in the making
• Machine Creativity: what it is and what it isn't
• BBC's Across the Board: Demis Hassabis
• The adventure of chess programming (3)
• MuZero figures out chess, rules and all
• BBC: Google DeepMind boss wins Nobel for proteins breakthrough