Offline Reinforcement Learning with On-Policy Q-Function Regularization

Abstract

The core challenge of offline reinforcement learning (RL) is dealing with the (potentially catastrophic) extrapolation error induced by the distribution shift between the history dataset and the desired policy. A large portion of prior work tackles this challenge by implicitly/explicitly regularizing the learning policy towards the behavior policy, which is hard to estimate reliably in practice. In this work, we propose to regularize towards the Q-function of the behavior policy instead of the behavior policy itself, under the premise that the Q-function can be estimated more reliably and easily by a SARSA-style estimate and handles the extrapolation error more straightforwardly. We propose two algorithms taking advantage of the estimated Q-function through regularizations, and demonstrate they exhibit strong performance on the D4RL benchmarks.

Cite

Text

Shi et al. "Offline Reinforcement Learning with On-Policy Q-Function Regularization." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2023. doi:10.1007/978-3-031-43421-1_27

Markdown

[Shi et al. "Offline Reinforcement Learning with On-Policy Q-Function Regularization." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2023.](https://mlanthology.org/ecmlpkdd/2023/shi2023ecmlpkdd-offline/) doi:10.1007/978-3-031-43421-1_27

BibTeX

@inproceedings{shi2023ecmlpkdd-offline,
  title     = {{Offline Reinforcement Learning with On-Policy Q-Function Regularization}},
  author    = {Shi, Laixi and Dadashi, Robert and Chi, Yuejie and Castro, Pablo Samuel and Geist, Matthieu},
  booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
  year      = {2023},
  pages     = {455-471},
  doi       = {10.1007/978-3-031-43421-1_27},
  url       = {https://mlanthology.org/ecmlpkdd/2023/shi2023ecmlpkdd-offline/}
}