DeepRead
Cover of Reinforcement Learning

Reinforcement Learning

Richard S. Sutton · 1992

Machine learning, statistics and probability › Deep learning and AI

#1,454 most recommended

About the book

A collection of seven invited research contributions edited by Richard S. Sutton, covering reinforcement learning: learning a mapping from situations to actions that maximizes a scalar reward signal. Its scope spans artificial intelligence, machine learning, control theory, and the psychology of learning that the field draws on. Reinforcement learning is distinguished from most machine learning by two features: trial-and-error search, since the learner must discover which actions pay off by trying them, and delayed reward, since actions affect later situations and therefore later rewards. The volume frames the field as both new and old, tracing its roots to Samuel's 1959 checkers player, which used temporal-difference learning, and to decades of psychological work on operant conditioning and secondary reinforcement. It treats much of the field as a reverse engineering of those learning processes.

Written by Claude from what it knows of this book.

What Hacker News says 60% recommend · 0% critical

051020102015202020262018 peak · 7
This bookAverage book in Deep learning and AI

In the commenters’ own words

What Open Library readers say

No Open Library reader has rated it yet.

  • 19 want to read
  • 0 reading
  • 1 have read

Open Library’s page for the book

What critics say, via Wikipedia

No Wikipedia article about this book yet, so no critics to quote.

Reddit communities — coming later

The threads it came up in

Details

First published
1992
ISBN
9780262193986
Editions known
18 at Open Library
Open Library
OL19899345W

Artificial intelligenceComputer scienceMachine learningPsychology ReinforcementReinforcement learning

Explore in the mosaic

Read next: its neighbours on the mosaic