Statistical Guarantees For Lifelong Reinforcement Learning Using Pac-bayes Theory

Abstract

Lifelong reinforcement learning (RL) has been developed as a paradigm for extending single-task RL to more realistic, dynamic settings. In lifelong RL, the "life" of an RL agent is modeled as a stream of tasks drawn from a task distribution. We propose EPIC (Empirical PAC-Bayes that Improves Continuously), a novel algorithm designed for lifelong RL using PAC-Bayes theory. EPIC learns a shared policy distribution, referred to as the world policy, which enables rapid adaptation to new tasks while retaining valuable knowledge from previous experiences. Our theoretical analysis establishes a relationship between the algorithm's generalization performance and the number of prior tasks preserved in memory. We also derive the sample complexity of EPIC in terms of RL regret. Extensive experiments on a variety of environments demonstrate that EPIC significantly outperforms existing methods in lifelong RL, offering both theoretical guarantees and practical efficacy through the use of the world p

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyzhang2024statistical

Related papers

• Lipschitz Lifelong Reinforcement Learning (2020)8.35
• Pac-bayesian Reinforcement Learning Trains Generalizable Policies (2025)0.00
• Deep Exploration With Pac-bayes (2024)0.00
• Lifelong Reinforcement Learning With Modulating Masks (2022)0.00
• Unifying PAC And Regret: Uniform PAC Bounds For Episodic Reinforcement Learning (2017)0.00
• Continuous Coordination As A Realistic Scenario For Lifelong Learning (2021)0.00
• Beyond No Regret: Instance-dependent PAC Reinforcement Learning (2021)0.00
• Reactive Exploration To Cope With Non-stationarity In Lifelong Reinforcement Learning (2022)0.00