A Kernel-Based Approach to Non-Stationary Reinforcement Learning in Metric Spaces

Abstract

In this work, we propose KeRNS: an algorithm for episodic reinforcement learning in non-stationary Markov Decision Processes (MDPs) whose state-action set is endowed with a metric. Using a non-parametric model of the MDP built with time-dependent kernels, we prove a regret bound that scales with the covering dimension of the state-action space and the total variation of the MDP with time, which quantifies its level of non-stationarity. Our method generalizes previous approaches based on sliding windows and exponential discounting used to handle changing environments. We further propose a practical implementation of KeRNS, we analyze its regret and validate it experimentally.

Cite

Text

Darwiche Domingues et al. "A Kernel-Based Approach to Non-Stationary Reinforcement Learning in Metric Spaces." Artificial Intelligence and Statistics, 2021.

Markdown

[Darwiche Domingues et al. "A Kernel-Based Approach to Non-Stationary Reinforcement Learning in Metric Spaces." Artificial Intelligence and Statistics, 2021.](https://mlanthology.org/aistats/2021/darwichedomingues2021aistats-kernelbased/)

BibTeX

@inproceedings{darwichedomingues2021aistats-kernelbased,
  title     = {{A Kernel-Based Approach to Non-Stationary Reinforcement Learning in Metric Spaces}},
  author    = {Darwiche Domingues, Omar and Menard, Pierre and Pirotta, Matteo and Kaufmann, Emilie and Valko, Michal},
  booktitle = {Artificial Intelligence and Statistics},
  year      = {2021},
  pages     = {3538-3546},
  volume    = {130},
  url       = {https://mlanthology.org/aistats/2021/darwichedomingues2021aistats-kernelbased/}
}