The same AI engine that made computers even better at games like Chess and Go could also boost new drug discoveries.
In 2016, AI company Deepmind, an Alphabet subsidiary, successfully did what no other software development company had done before: It had created a program, AlphaGo, that could successfully defeat a world champion in the complex, ancient game of Go. Less than two years later, its program AlphaZero showed it could play at a higher level than any other existing chess program (which had left humans in the dust years before that.)
But these programs didn’t work through mere brute force calculation: They used sophisticated machine learning algorithms that not only performed better than previous computer game engines, helped they unveil new insights about the games of chess and Go themselves. Since these famous matches, both games have seen a burst of creativity inspired by some of the new strategic insights the algorithms provided.
“Playing games like chess has been a huge part of my life since childhood,” says Deepmind cofounder and CEO Demis Hassibis. “It felt to me like a natural platform to develop and test our algorithmic ideas.”
Since those initial gaming revolutions, Deepmind went on to develop a new machine learning engine, called AlphaFold, to solve a deep and mysterious problem in an entirely different sphere: There are an exponentially staggering number of ways for proteins to potentially fold themselves. So how do they quickly, consistently fold themselves in the same way? Solving this biological puzzle would enormously benefit the development and treatment of disease as well as how biology generally operates. But for decades it was thought the amount of computational power needed was well beyond what was available.
However, using the experience of designing engines that can learn complicated games like chess in a matter of hours, Deepmind researcher John Jumper led a team that developed AlphaFold, which was able to successfully predict the way proteins fold. This led to a major paper in Nature and has already helped stimulate new pathways of drug discovery. It’s also unleashed the creativity of biologists, says Jumper. “I see AlphaFold as supporting the creativity of biologists and letting them ask some of these ‘what if’ questions extremely quickly,” he says.
On Thursday, Hemis and Jumper won one of the five $3 million Breakthrough Prizes for 2023 for the development of AlphaFold. The honor was shared among 11 different scientists working in Fundamental Physics, Life Sciences and Mathematics. Additionally, the Breakthrough Prizes also awarded three $100,000 New Horizons in Mathematics prizes, three $100,000 New Horizons Prizes for Physics and three $50,000 Maryam Mirzakhani New Frontiers Prizes. The New Horizons awards honor work by early-career scientists and the Maryam Mirzakhani award is for women working in mathematics. A total of $15,750,000 in prizes are being awarded.
This money is granted by the Breakthrough Prize Foundation, which was founded by billionaire Yuri Milner and his wife Julia. The prizes themselves were cofounded by the Milners, 23andMe CEO Anne Wojcicki, Google cofounder Sergey Brin, Facebook CEO Mark Zuckerberg and his wife Priscilla Chan.
Breakthrough Prizes In Life Sciences
Two other prizes in Life Sciences were awarded besides the one granted to Hemis and Jumper. First, there was a $3 million prize awarded to Clifford Brangwynne and Anthony Hyman. These two researchers discovered that cell infrastructure, genetic regulation and other processes sometimes occur between cells as a result of rapidly forming liquid droplets. This discovery has helped deepen understanding of processes that occur within organisms, and may hold a key to treating certain degenerative diseases.
Another set of $3 million Life Sciences Award Winners were Masashi Yanagisawa and Emmanuel Mignot. These two researchers, pursuing independent lines of inquiry, helped uncover one of the central causes of narcolepsy. These discoveries helped pave the way for treating the disorder. And in addition to that, the molecular pathways discovered also led to the development of new ways to help people with sleep disorders rest easier.
Breakthrough Prize in Mathematics
Yale mathematician Daniel Spielman was awarded $3 million for his discoveries in computer science and mathematics. His work has helped develop better error correction in computation, and algorithms he helped discover are used in a wide variety of applications in computing and engineering. Currently, he’s also engaged in the application of math to design better clinical trials for drug development. Additionally, he and collaborators also successfully solved a quantum mechanics problem called the Kadison-Singer problem. In the process, that mathematical solution has turned out to have unexpected applications in pure mathematics as well as optimization problems that are key to manufacturing and supply chain processes.
Breakthrough Prize In Fundamental Physics
A $3 million prize for fundamental physics was awarded to four people: Charles Bennett, Gilles Brassard, Peter Shor and David Deutsch. All of these people are responsible for discovering the fundamental physics that underlie quantum computation. Over the past few years, there’s been a surge of interest in quantum computing, with millions being poured into the field by both investors backing startups as well as large companies like Google, Microsoft and IBM, who hope to gain an edge in the quantum computing space. The technology promises to revolutionize difficult problems such as simulating chemical reactions, solving complicated logistics calculations for supply chains and shipping and machine learning algorithms.
“I think I was more pleased that nobody will think it’s strange that people are winning prizes for quantum computing,” Deutsch told Forbes when asked about how it felt to be awarded the prize. “It’s sort of entered the mainstream of things that people know exist. And, you know, I remember a time when there were only a handful of people that knew it existed!”