English

Reward Centering

Machine Learning 2024-10-31 v2 Artificial Intelligence

Abstract

We show that discounted methods for solving continuing reinforcement learning problems can perform significantly better if they center their rewards by subtracting out the rewards' empirical average. The improvement is substantial at commonly used discount factors and increases further as the discount factor approaches one. In addition, we show that if a problem's rewards are shifted by a constant, then standard methods perform much worse, whereas methods with reward centering are unaffected. Estimating the average reward is straightforward in the on-policy setting; we propose a slightly more sophisticated method for the off-policy setting. Reward centering is a general idea, so we expect almost every reinforcement-learning algorithm to benefit by the addition of reward centering.

Keywords

Cite

@article{arxiv.2405.09999,
  title  = {Reward Centering},
  author = {Abhishek Naik and Yi Wan and Manan Tomar and Richard S. Sutton},
  journal= {arXiv preprint arXiv:2405.09999},
  year   = {2024}
}

Comments

In Proceedings of RLC 2024

R2 v1 2026-06-28T16:29:20.598Z