A Case for Validation Buffer in Pessimistic Actor-Critic

2024/03/01 by Michał Nauman, Mateusz Ostaszewski, Nauman, Michal +3
Computer Science · Neuroscience · #Chaos, Complexity, and Education #Cognitive Science and Education Research #FOS: Computer and information sciences #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2403.01014

openalex publication_date 2024/03/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

In this paper, we investigate the issue of error accumulation in critic networks updated via pessimistic temporal difference objectives. We show that the critic approximation error can be approximated via a recursive fixed-point model similar to that of the Bellman value. We use such recursive definition to retrieve the conditions under which the pessimistic critic is unbiased. Building on these insights, we propose Validation Pessimism Learning (VPL) algorithm. VPL uses a small validation buffer to adjust the levels of pessimism throughout the agent training, with the pessimism set such that the approximation error of the critic targets is minimized. We investigate the proposed approach on a variety of locomotion and manipulation tasks and report improvements in sample efficiency and performance.

Related

• Improving Actor-Critic Training with Steerable Action-Value Approximation Errors
• Tactical Optimism and Pessimism for Deep Reinforcement Learning
• Bellman-consistent Pessimism for Offline Reinforcement Learning
• Learning Pessimism for Robust and Efficient Off-Policy Reinforcement Learning
• Provable Benefits of Actor-Critic Methods for Offline Reinforcement Learning