Momentum in Reinforcement Learning
Abstract
We adapt the optimization’s concept of momentum to reinforcement learning. Seeing the state-action value functions as an anlog to the gradients in optimization, we interpret momentum as an average of consecutive $q$-functions. We derive Momentum Value Iteration (MoVI), a variation of Value iteration that incorporates this momentum idea. Our analysis shows that this allows MoVI to average errors over successive iterations. We show that the proposed approach can be readily extended to deep learning. Specifically,we propose a simple improvement on DQN based on MoVI, and experiment it on Atari games.
Cite
Text
Vieillard et al. "Momentum in Reinforcement Learning." Artificial Intelligence and Statistics, 2020.Markdown
[Vieillard et al. "Momentum in Reinforcement Learning." Artificial Intelligence and Statistics, 2020.](https://mlanthology.org/aistats/2020/vieillard2020aistats-momentum/)BibTeX
@inproceedings{vieillard2020aistats-momentum,
title = {{Momentum in Reinforcement Learning}},
author = {Vieillard, Nino and Scherrer, Bruno and Pietquin, Olivier and Geist, Matthieu},
booktitle = {Artificial Intelligence and Statistics},
year = {2020},
pages = {2529-2538},
volume = {108},
url = {https://mlanthology.org/aistats/2020/vieillard2020aistats-momentum/}
}