Better Exploration with Optimistic Actor-Critic

TL;DR

This paper introduces Optimistic Actor Critic, a new reinforcement learning algorithm that improves exploration efficiency by using confidence bounds, leading to state-of-the-art performance in continuous control tasks.

Contribution

The paper proposes a novel algorithm, Optimistic Actor Critic, which incorporates confidence bounds to enhance exploration in actor-critic methods, addressing key limitations of existing algorithms.

Findings

Achieves state-of-the-art sample efficiency in continuous control tasks.

Addresses pessimistic underexploration by using confidence bounds.

Demonstrates improved exploration through directed action sampling.

Abstract

Actor-critic methods, a type of model-free Reinforcement Learning, have been successfully applied to challenging tasks in continuous control, often achieving state-of-the art performance. However, wide-scale adoption of these methods in real-world domains is made difficult by their poor sample efficiency. We address this problem both theoretically and empirically. On the theoretical side, we identify two phenomena preventing efficient exploration in existing state-of-the-art algorithms such as Soft Actor Critic. First, combining a greedy actor update with a pessimistic estimate of the critic leads to the avoidance of actions that the agent does not know about, a phenomenon we call pessimistic underexploration. Second, current algorithms are directionally uninformed, sampling actions with equal probability in opposite directions from the current mean. This is wasteful, since we typically need actions taken along certain directions much more than others. To address both of these phenomena, we introduce a new algorithm, Optimistic Actor Critic, which approximates a lower and upper confidence bound on the state-action value function. This allows us to apply the principle of optimism in the face of uncertainty to perform directed exploration using the upper bound while still using the lower bound to avoid overestimation. We evaluate OAC in several challenging continuous control tasks, achieving state-of the art sample efficiency.