Regularized Anderson Acceleration for Off-Policy Deep Reinforcement Learning

TL;DR

This paper introduces a regularized Anderson acceleration method to enhance the convergence speed and performance of off-policy deep reinforcement learning algorithms, effectively addressing sample inefficiency and slow learning in complex environments.

Contribution

It extends Anderson acceleration with regularization to deep RL, proposing strategies like progressive update and adaptive restart to improve learning efficiency.

Findings

Significantly faster learning speed on benchmark tasks

Improved final performance of deep RL algorithms

Effective in high-dimensional, continuous control environments

Abstract

Model-free deep reinforcement learning (RL) algorithms have been widely used for a range of complex control tasks. However, slow convergence and sample inefficiency remain challenging problems in RL, especially when handling continuous and high-dimensional state spaces. To tackle this problem, we propose a general acceleration method for model-free, off-policy deep RL algorithms by drawing the idea underlying regularized Anderson acceleration (RAA), which is an effective approach to accelerating the solving of fixed point problems with perturbations. Specifically, we first explain how policy iteration can be applied directly with Anderson acceleration. Then we extend RAA to the case of deep RL by introducing a regularization term to control the impact of perturbation induced by function approximation errors. We further propose two strategies, i.e., progressive update and adaptive restart, to enhance the performance. The effectiveness of our method is evaluated on a variety of benchmark tasks, including Atari 2600 and MuJoCo. Experimental results show that our approach substantially improves both the learning speed and final performance of state-of-the-art deep RL algorithms.