The asymptotic equipartition property in reinforcement learning and its relation to return maximization

DOI10.1016/j.neunet.2005.02.008zbMath1093.68082OpenAlexW2171630474WikidataQ51962962 ScholiaQ51962962MaRDI QIDQ2488678

Hideaki Sakai, Kazushi Ikeda, Kazunori Iwata

Publication date: 11 May 2006

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

Full work available at URL: https://doi.org/10.1016/j.neunet.2005.02.008

zbMATH Keywords

Markov decision processinformation theoryreinforcement learningstochastic complexityasymptotic equipartition propertyreturn maximization

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Related Items (1)

An information-theoretic analysis of return maximization in reinforcement learning

Cites Work

• A Mathematical Theory of Communication
• Asynchronous stochastic approximation and Q-learning
• Convergence results for single-step on-policy reinforcement-learning algorithms
• \({\mathcal Q}\)-learning
• The convergence of \(TD(\lambda)\) for general \(\lambda\)
• Universal coding with minimum probability of codeword length overflow
• Algorithmic Information Theory
• Statistical inference under multiterminal rate restrictions: a differential geometric approach
• The error exponent for the noiseless encoding of finite ergodic Markov sources
• Variable-to-fixed length codes provide better large deviations performance than fixed-to-variable length codes
• Algorithmic Information Theory
• Reliability function of a discrete memoryless channel at rates above capacity (Corresp.)
• On the Convergence of Stochastic Iterative Dynamic Programming Algorithms
• The method of types [information theory]
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item

This page was built for publication: The asymptotic equipartition property in reinforcement learning and its relation to return maximization