Maximum Entropy RL (Provably) Solves Some Robust RL Problems

TL;DR

This paper provides a theoretical proof that maximum entropy reinforcement learning inherently maximizes a lower bound on a robust RL objective, demonstrating its robustness to certain disturbances without extra modifications.

Contribution

The work offers the first rigorous proof and theoretical characterization of MaxEnt RL's robustness to disturbances in dynamics and reward functions.

Findings

MaxEnt RL maximizes a lower bound on a robust RL objective.

MaxEnt RL is robust to certain disturbances without additional modifications.

Provides formal guarantees for MaxEnt RL's robustness.

Abstract

Many potential applications of reinforcement learning (RL) require guarantees that the agent will perform well in the face of disturbances to the dynamics or reward function. In this paper, we prove theoretically that maximum entropy (MaxEnt) RL maximizes a lower bound on a robust RL objective, and thus can be used to learn policies that are robust to some disturbances in the dynamics and the reward function. While this capability of MaxEnt RL has been observed empirically in prior work, to the best of our knowledge our work provides the first rigorous proof and theoretical characterization of the MaxEnt RL robust set. While a number of prior robust RL algorithms have been designed to handle similar disturbances to the reward function or dynamics, these methods typically require additional moving parts and hyperparameters on top of a base RL algorithm. In contrast, our results suggest that MaxEnt RL by itself is robust to certain disturbances, without requiring any additional modifications. While this does not imply that MaxEnt RL is the best available robust RL method, MaxEnt RL is a simple robust RL method with appealing formal guarantees.