English

Offline Reinforcement Learning Hands-On

Machine Learning 2020-12-01 v1

Abstract

Offline Reinforcement Learning (RL) aims to turn large datasets into powerful decision-making engines without any online interactions with the environment. This great promise has motivated a large amount of research that hopes to replicate the success RL has experienced in simulation settings. This work ambitions to reflect upon these efforts from a practitioner viewpoint. We start by discussing the dataset properties that we hypothesise can characterise the type of offline methods that will be the most successful. We then verify these claims through a set of experiments and designed datasets generated from environments with both discrete and continuous action spaces. We experimentally validate that diversity and high-return examples in the data are crucial to the success of offline RL and show that behavioural cloning remains a strong contender compared to its contemporaries. Overall, this work stands as a tutorial to help people build their intuition on today's offline RL methods and their applicability.

Keywords

Cite

@article{arxiv.2011.14379,
  title  = {Offline Reinforcement Learning Hands-On},
  author = {Louis Monier and Jakub Kmec and Alexandre Laterre and Thomas Pierrot and Valentin Courgeau and Olivier Sigaud and Karim Beguir},
  journal= {arXiv preprint arXiv:2011.14379},
  year   = {2020}
}

Comments

Accepted at NeurIPS 2020 Offline Reinforcement Learning Workshop. First two authors contributed equally. Authors three and four advised equally

R2 v1 2026-06-23T20:34:46.678Z