Textbook of AI

1957–, Computer scientist

Richard Stuart Sutton is a Canadian computer scientist who, more than any other person, founded modern reinforcement learning. His 1984 PhD thesis with Andrew Barto introduced temporal-difference (TD) learning, a learning rule that updates a value estimate by the difference between successive predictions, generalising both the Monte Carlo and dynamic-programming approaches to value estimation. The 1988 paper Learning to Predict by the Methods of Temporal Differences gave TD learning its modern theoretical analysis.

With Barto, Sutton wrote Reinforcement Learning: An Introduction (1998, second edition 2018), the standard textbook of the field. His 2019 essay The Bitter Lesson, that "general methods that leverage computation are ultimately the most effective", has become a touchstone of the modern AI scaling era. He is professor at the University of Alberta and Chief Scientific Advisor at the Alberta Machine Intelligence Institute (Amii); from 2017 to early 2023 he led DeepMind's Edmonton lab (which DeepMind closed in 2023), and in 2024 he co-founded the AGI startup Keen Technologies with John Carmack. He shares (with Barto) the 2024 Turing Award for foundational contributions to reinforcement learning.