Classification-based Approximate Policy Iteration: Experiments and Extended Discussions
This work addresses scalability issues in reinforcement learning for researchers and practitioners, offering a flexible method that is incremental in combining existing techniques.
The paper tackles large approximate dynamic programming or reinforcement learning problems by introducing a classification-based approximate policy iteration (CAPI) framework that exploits regularities in both value functions and policy spaces, achieving tighter theoretical bounds on performance loss and demonstrating empirical results on tasks like HIV control.
Tackling large approximate dynamic programming or reinforcement learning problems requires methods that can exploit regularities, or intrinsic structure, of the problem in hand. Most current methods are geared towards exploiting the regularities of either the value function or the policy. We introduce a general classification-based approximate policy iteration (CAPI) framework, which encompasses a large class of algorithms that can exploit regularities of both the value function and the policy space, depending on what is advantageous. This framework has two main components: a generic value function estimator and a classifier that learns a policy based on the estimated value function. We establish theoretical guarantees for the sample complexity of CAPI-style algorithms, which allow the policy evaluation step to be performed by a wide variety of algorithms (including temporal-difference-style methods), and can handle nonparametric representations of policies. Our bounds on the estimation error of the performance loss are tighter than existing results. We also illustrate this approach empirically on several problems, including a large HIV control task.