English

Adaptive Lambda Least-Squares Temporal Difference Learning

Machine Learning 2017-01-02 v1 Artificial Intelligence Machine Learning

Abstract

Temporal Difference learning or TD(λ\lambda) is a fundamental algorithm in the field of reinforcement learning. However, setting TD's λ\lambda parameter, which controls the timescale of TD updates, is generally left up to the practitioner. We formalize the λ\lambda selection problem as a bias-variance trade-off where the solution is the value of λ\lambda that leads to the smallest Mean Squared Value Error (MSVE). To solve this trade-off we suggest applying Leave-One-Trajectory-Out Cross-Validation (LOTO-CV) to search the space of λ\lambda values. Unfortunately, this approach is too computationally expensive for most practical applications. For Least Squares TD (LSTD) we show that LOTO-CV can be implemented efficiently to automatically tune λ\lambda and apply function optimization methods to efficiently search the space of λ\lambda values. The resulting algorithm, ALLSTD, is parameter free and our experiments demonstrate that ALLSTD is significantly computationally faster than the na\"{i}ve LOTO-CV implementation while achieving similar performance.

Keywords

Cite

@article{arxiv.1612.09465,
  title  = {Adaptive Lambda Least-Squares Temporal Difference Learning},
  author = {Timothy A. Mann and Hugo Penedones and Shie Mannor and Todd Hester},
  journal= {arXiv preprint arXiv:1612.09465},
  year   = {2017}
}