Two trailblazing pc scientists have received the 2024 Turing Award for his or her work in reinforcement studying, a self-discipline wherein machines study by a reward-based trial-and-error strategy that lets them adapt inside constrained or dynamic environments.
Andrew G. Barto, a professor emeritus on the College of Massachusetts Amherst; and Richard S. Sutton, a professor on the College of Alberta, developed key algorithms and theories by a seminal sequence of papers starting in the 1980s. This consists of work on a reinforcement approach referred to as temporal difference learning; the duo later printed a tutorial textbook referred to as Reinforcement Learning: An Introduction.
Esteemed mathematician Alan Turing (pictured above), after whom the Turing Award is called, additionally produced a paper within the Nineteen Fifties referred to as Computing Machinery and Intelligence that questioned whether or not computer systems can suppose and touched on related ideas round studying from expertise.
In newer years, reinforcement studying has obtained extra consideration after Google Deepmind used the approach to construct an AI that defeated the world’s greatest AlphaGo gamers. And prior to now few months, Chinese language AI upstart DeepSeek hit the headlines for its game-changing R1 reasoning mannequin, which leaned closely on reinforcement studying to create cheaper basis fashions.
‘Nobel Prize for computing’
The Turing Award, administered by the Affiliation for Computing Equipment (ACM), has typically been dubbed the “Nobel Prize for computing.” Nevertheless, the Nobel Prize itself has been encroaching into the computing realm, significantly round AI; Geoff Hinton and John Hopfield received the Nobel Prize in Physics for his or her work in foundational AI final yr. This was adopted shortly after by DeepMind’s Demis Hassabis and John Jumper who have been awarded the Nobel Prize in Chemistry for his or her work on AlphaFold.
“Analysis areas starting from cognitive science and psychology to neuroscience impressed the event of reinforcement studying, which has laid the foundations for among the most vital advances in AI and has given us higher perception into how the mind works,” ACM president Yannis Ioannidis stated in a press release. “Barto and Sutton’s work shouldn’t be a stepping stone that we now have now moved on from. Reinforcement studying continues to develop and gives nice potential for additional advances in computing and lots of different disciplines. It’s becoming that we’re honoring them with essentially the most prestigious award in our subject.”
Different notable AI pioneers to win the Turing Award embody Meta’s chief AI scientist Yann LeCun, who was awarded the prize in 2018 alongside Geoff Hinton and Yoshua Bengio for his or her work on deep neural networks.
Barto and Sutton will share the $1 million money prize, which was supplied with help from Google.