Reinforcement Learning
by Richard S. Sutton
Richard Sutton's definitive textbook on agents maximizing reward through interaction with uncertain environments. The gold standard RL text.
"Reinforcement learning is simultaneously a problem, a class of solution methods, and the field that studies both".
Editorial Summary
Reinforcement learning is a computational approach to learning whereby an agent tries to maximize the total amount of reward it receives while interacting with a complex, uncertain environment. In Reinforcement Learning, Richard Sutton and Andrew Barto provide a clear and simple account of the field's key ideas and algorithms. Richard S. Sutton is Professor of Computing Science and AITF Chair in Reinforcement Learning and Artificial Intelligence at the University of Alberta, and also Distinguished Research Scientist at DeepMind. This second edition has been significantly expanded and updated, presenting new topics and updating coverage of other topics. Many algorithms presented in this part are new to the second edition, including UCB, Expected Sarsa, and Double Learning.
Perspective
"Reading this feels like getting a master class from the field's founding father, as Sutton methodically builds your intuition from simple bandits to complex policy gradients, making you think like an agent optimizing for long-term reward. The book's distinctive contribution is formalizing reinforcement learning as optimal control of incompletely-known Markov decision processes, providing the mathematical foundation that unified scattered approaches into a coherent field. Machine learning engineers building autonomous systems need this theoretical grounding now more than ever, as they will find the conceptual framework to understand why their deep reinforcement learning algorithms work."
Matched by concept and theme
