Reinforcement Learning under Model Mismatch
Abstract
We study reinforcement learning under model misspecification, where we do not have access to the true environment but only to a reasonably close approximation to it. We address this problem by extending the framework of robust MDPs to the modelfree Reinforcement Learning setting, where we do not have access to the model parameters, but can only sample states from it. We define robust versions of Qlearning, SARSA, and TDlearning and prove convergence to an approximately optimal robust policy and approximate value function respectively. We scale up the robust algorithms to large MDPs via function approximation and prove convergence under two different settings. We prove convergence of robust approximate policy iteration and robust approximate value iteration for linear architectures (under mild assumptions). We also define a robust loss function, the mean squared robust projected Bellman error and give stochastic gradient descent algorithms that are guaranteed to converge to a local minimum.
 Publication:

arXiv eprints
 Pub Date:
 June 2017
 arXiv:
 arXiv:1706.04711
 Bibcode:
 2017arXiv170604711R
 Keywords:

 Computer Science  Machine Learning;
 Statistics  Machine Learning
 EPrint:
 To appear in Proceedings of NIPS 2017