Statistical Guarantees for Lifelong Reinforcement Learning using PAC-Bayes Theory
This work addresses the challenge of dynamic, multi-task learning in reinforcement learning, providing both theoretical guarantees and practical benefits for AI agents in evolving environments.
The paper tackles the problem of lifelong reinforcement learning by proposing EPIC, a novel algorithm that learns a shared policy distribution for rapid adaptation to new tasks, and demonstrates significant performance improvements over existing methods in experiments.
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 policy.