English

Efficient Wasserstein Natural Gradients for Reinforcement Learning

Machine Learning 2021-03-19 v4

Abstract

A novel optimization approach is proposed for application to policy gradient methods and evolution strategies for reinforcement learning (RL). The procedure uses a computationally efficient Wasserstein natural gradient (WNG) descent that takes advantage of the geometry induced by a Wasserstein penalty to speed optimization. This method follows the recent theme in RL of including a divergence penalty in the objective to establish a trust region. Experiments on challenging tasks demonstrate improvements in both computational cost and performance over advanced baselines.

Keywords

Cite

@article{arxiv.2010.05380,
  title  = {Efficient Wasserstein Natural Gradients for Reinforcement Learning},
  author = {Ted Moskovitz and Michael Arbel and Ferenc Huszar and Arthur Gretton},
  journal= {arXiv preprint arXiv:2010.05380},
  year   = {2021}
}
R2 v1 2026-06-23T19:15:35.130Z