Provably Convergent Two-Timescale Off-Policy Actor-Critic with Function Approximation

2019/11/11 by Shangtong Zhang, Bo Liu, Zhang, Shangtong +5 · 6 citations
Computer Science · Physics and Astronomy · #Adaptive Dynamic Programming Control #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Model Reduction and Neural Networks #Reinforcement Learning in Robotics

paper · pdf · doi:10.48550/arxiv.1911.04384

openalex publication_date 2019/11/11 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We present the first provably convergent two-timescale off-policy actor-critic algorithm (COF-PAC) with function approximation. Key to COF-PAC is the introduction of a new critic, the emphasis critic, which is trained via Gradient Emphasis Learning (GEM), a novel combination of the key ideas of Gradient Temporal Difference Learning and Emphatic Temporal Difference Learning. With the help of the emphasis critic and the canonical value function critic, we show convergence for COF-PAC, where the critics are linear and the actor can be nonlinear.

Cited by

• Average-Reward Off-Policy Policy Evaluation with Function Approximation
• Truncated Emphatic Temporal Difference Methods for Prediction and Control
• PER-ETD: A Polynomially Efficient Emphatic Temporal Difference Learning Method
• Breaking the Deadly Triad with a Target Network
• Finite-Sample Analysis of Off-Policy Natural Actor-Critic Algorithm
• Mean-Variance Policy Iteration for Risk-Averse Reinforcement Learning

Related

• Single-Timescale Actor-Critic Provably Finds Globally Optimal Policy
• Convergence Proof for Actor-Critic Methods Applied to PPO and RUDDER
• Convergent Actor-Critic Algorithms Under Off-Policy Training and Function Approximation
• A Finite Time Analysis of Two Time-Scale Actor Critic Methods
• Non-asymptotic Convergence Analysis of Two Time-scale (Natural) Actor-Critic Algorithms