An information-theoretic analysis of return maximization in reinforcement learning

DOI10.1016/J.NEUNET.2011.05.002zbMATH Open1266.68156OpenAlexW2034994237WikidataQ51559078 ScholiaQ51559078MaRDI QIDQ2375396FDOQ2375396

Authors: Kazunori Iwata 

Publication date: 14 June 2013

Published in: Neural Networks (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1016/j.neunet.2011.05.002

Recommendations

• The asymptotic equipartition property in reinforcement learning and its relation to return maximization
• An information-theoretic analysis of Thompson sampling
• Reinforcement learning in finite MDPs: PAC analysis
• Near-optimal reinforcement learning in polynomial time
• Using Expectation-Maximization for Reinforcement Learning

zbMATH Keywords

information theoryreinforcement learningasymptotic equipartition propertystochastic sequential decision process

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Cites Work

• Elements of Information Theory
• Title not available (Why is that?)
• Title not available (Why is that?)
• A Mathematical Theory of Communication
• Title not available (Why is that?)
• \({\mathcal Q}\)-learning
• Title not available (Why is that?)
• Approximation theory of output statistics
• Title not available (Why is that?)
• The Basic Theorems of Information Theory
• The Individual Ergodic Theorem of Information Theory
• Title not available (Why is that?)
• The convergence of \(TD(\lambda)\) for general \(\lambda\)
• Discrete Dynamic Programming
• Asymptotically mean stationary measures
• The method of types [information theory]
• Asynchronous stochastic approximation and Q-learning
• The O.D.E. Method for Convergence of Stochastic Approximation and Reinforcement Learning
• The strong ergodic theorem for densities: Generalized Shannon-McMillan- Breiman theorem
• A simple proof of the Moy-Perez generalization of the Shannon-McMillan theorem
• A New Optimality Criterion for Nonhomogeneous Markov Decision Processes
• Convergence results for single-step on-policy reinforcement-learning algorithms
• Correction Notes: Correction to "The Individual Ergodic Theorem of Information Theory"
• Generalizations of Shannon-McMillan theorem
• Boundedness of iterates in \(Q\)-learning
• The asymptotic equipartition property in reinforcement learning and its relation to return maximization
• The role of the asymptotic equipartition property in noiseless source coding
• Model-Based Reinforcement Learning for Partially Observable Games with Sampling-Based State Estimation

Cited In (3)

• The asymptotic equipartition property in reinforcement learning and its relation to return maximization
• An information-theoretic analysis of Thompson sampling
• Trading utility and uncertainty: applying the value of information to resolve the exploration-exploitation dilemma in reinforcement learning

This page was built for publication: An information-theoretic analysis of return maximization in reinforcement learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2375396)