Andrew Barto, Professor Emeritus of Information and Computer Sciences  at the University of Massachusetts, Amherst and Richard Sutton, a professor of Computer Science at the University of Alberta and former DeepMind researcher, are the joint recipients of the 2024 Turing Award for developing the conceptual and algorithmic foundations of reinforcement learning.

Named  after Alan Turning this is the premier award for computing made by the ACM  and carries a $1 million prize with financial support from Google.

It is particularly timely that the Turing Award for 2024 should be made in respect of an aspect of AI that Turing himself had experimented with. As the ACM announcement reminds us,  Alan Turing’s 1950 paper “Computing Machinery and Intelligence,” addressed the question “Can machines think?” and proposed an approach to machine learning based on rewards and punishments, i.e the mechanism of reinforcement learning.

Inspired by psychological studies of human learning, Andrew Barto and Richard Sutton, who was his PhD student, began work on what they termed "reinforcement learning" in the 1980s,  and continued to work on it during the period often referred to as the second AI winter, publishing their textbook "Reinforcement Learning: An Introduction" in 1998. This book was very influential and a revised and updated edition was published in 2018 and is included in our Book Watch Archive, see Reinforcement Learning: An Introduction 2nd Ed (Bradford Books).  

It is also apt that the prize money comes from Google since the breakthrough in using reinforcement learning, as distinct from supervised learning and the use of large datasets, was made by Google with AlphaGo which taught itself how to play Go - and later how to do protein folding and all the other remarkable skills it has mastered - by the intrinsic reward of success.

Following in the footsteps of Geoffrey Hinton who used winning the Nobel Prize as an opportunity to talk about the dangers of the rapid advances in AI,  Andrew Barto took the opportunity to raise concerns warning  about the unsafe deployment of AI systems, saying:

 "Releasing software to millions of people without safeguards is not good engineering practice comparing it to testing a bridge by having people use it."

Both laureates also criticized President Trump's proposed cuts to federal research funding, with Barto calling it a "tragedy" that would eliminate opportunities for exploratory research like their early work. 

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info

<ASIN:B014I5YDNM>

<ASIN:0262039249>