English

Reinforcement Learning using Guided Observability

Machine Learning 2021-04-23 v1

Abstract

Due to recent breakthroughs, reinforcement learning (RL) has demonstrated impressive performance in challenging sequential decision-making problems. However, an open question is how to make RL cope with partial observability which is prevalent in many real-world problems. Contrary to contemporary RL approaches, which focus mostly on improved memory representations or strong assumptions about the type of partial observability, we propose a simple but efficient approach that can be applied together with a wide variety of RL methods. Our main insight is that smoothly transitioning from full observability to partial observability during the training process yields a high performance policy. The approach, called partially observable guided reinforcement learning (PO-GRL), allows to utilize full state information during policy optimization without compromising the optimality of the final policy. A comprehensive evaluation in discrete partially observableMarkov decision process (POMDP) benchmark problems and continuous partially observable MuJoCo and OpenAI gym tasks shows that PO-GRL improves performance. Finally, we demonstrate PO-GRL in the ball-in-the-cup task on a real Barrett WAM robot under partial observability.

Keywords

Cite

@article{arxiv.2104.10986,
  title  = {Reinforcement Learning using Guided Observability},
  author = {Stephan Weigand and Pascal Klink and Jan Peters and Joni Pajarinen},
  journal= {arXiv preprint arXiv:2104.10986},
  year   = {2021}
}
R2 v1 2026-06-24T01:25:38.789Z