Header menu link for other important links
X
Learning to repeat: Fine grained action repetition for deep reinforcement learning
Aravind S. Lakshminarayanan, Balaraman Ravindran
Published in International Conference on Learning Representations, ICLR
2019
Abstract
Reinforcement Learning algorithms can learn complex behavioral patterns for sequential decision making tasks wherein an agent interacts with an environment and acquires feedback in the form of rewards sampled from it. Traditionally, such algorithms make decisions, i.e., select actions to execute, at every single time step of the agent-environment interactions. In this paper, we propose a novel framework, Fine Grained Action Repetition (FiGAR), which enables the agent to decide the action as well as the time scale of repeating it. FiGAR can be used for improving any Deep Reinforcement Learning algorithm which maintains an explicit policy estimate by enabling temporal abstractions in the action space. We empirically demonstrate the efficacy of our framework by showing performance improvements on top of three policy search algorithms in different domains: Asynchronous Advantage Actor Critic in the Atari 2600 domain, Trust Region Policy Optimization in Mujoco domain and Deep Deterministic Policy Gradients in the TORCS car racing domain. © ICLR 2019 - Conference Track Proceedings. All rights reserved.
About the journal
Journal5th International Conference on Learning Representations, ICLR 2017 - Conference Track Proceedings
PublisherInternational Conference on Learning Representations, ICLR
Open AccessNo
Concepts (14)
  •  related image
    Decision making
  •  related image
    Learning algorithms
  •  related image
    Machine learning
  •  related image
    Reinforcement learning
  •  related image
    Software agents
  •  related image
    ACTION SPACES
  •  related image
    Behavioral patterns
  •  related image
    Different domains
  •  related image
    POLICY GRADIENT
  •  related image
    POLICY OPTIMIZATION
  •  related image
    POLICY SEARCH
  •  related image
    Sequential decision making
  •  related image
    TEMPORAL ABSTRACTION
  •  related image
    Deep learning