Corpus ID: 218665668

Think Too Fast Nor Too Slow: The Computational Trade-off Between Planning And Reinforcement Learning

  title={Think Too Fast Nor Too Slow: The Computational Trade-off Between Planning And Reinforcement Learning},
  author={T. M. Moerland and Anna Deichler and S. Baldi and J. Broekens and Catholijn M. Jonker},
  • T. M. Moerland, Anna Deichler, +2 authors Catholijn M. Jonker
  • Published 2020
  • Computer Science
  • ArXiv
  • Planning and reinforcement learning are two key approaches to sequential decision making. Multi-step approximate real-time dynamic programming, a recently successful algorithm class of which AlphaZero [Silver et al., 2018] is an example, combines both by nesting planning within a learning loop. However, the combination of planning and learning introduces a new question: how should we balance time spend on planning, learning and acting? The importance of this trade-off has not been explicitly… CONTINUE READING
    3 Citations
    Model-based Reinforcement Learning: A Survey
    • 3
    • PDF
    A Framework for Reinforcement Learning and Planning
    • 5
    • PDF
    CAMPs: Learning Context-Specific Abstractions for Efficient Planning in Factored MDPs
    • PDF


    Thinking Fast and Slow with Deep Learning and Tree Search
    • 143
    • PDF
    Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
    • 314
    • PDF
    Learning to Act Using Real-Time Dynamic Programming
    • 1,259
    • PDF
    Reinforcement Learning: An Introduction
    • 26,976
    • Highly Influential
    • PDF
    Multi-Step Greedy and Approximate Real Time Dynamic Programming
    • 4
    Benchmarking Model-Based Reinforcement Learning
    • 104
    • PDF
    Dyna, an integrated architecture for learning, planning, and reacting
    • 524
    Mastering the game of Go without human knowledge
    • 3,773
    • Highly Influential
    • PDF
    Exploration Strategies for Model-based Learning in Multi-agent Systems: Exploration Strategies
    • 62
    • PDF