Learning Robust Options
This work addresses robust reinforcement learning for temporally abstract actions, offering incremental improvements in handling model uncertainty for tasks with uncertain environments.
The paper tackles the problem of learning temporally abstract actions robust to model uncertainty in reinforcement learning, proposing a Robust Options Policy Iteration algorithm with convergence guarantees and demonstrating that robustness improves generalization over a broader range of dynamics compared to non-robust methods.
Robust reinforcement learning aims to produce policies that have strong guarantees even in the face of environments/transition models whose parameters have strong uncertainty. Existing work uses value-based methods and the usual primitive action setting. In this paper, we propose robust methods for learning temporally abstract actions, in the framework of options. We present a Robust Options Policy Iteration (ROPI) algorithm with convergence guarantees, which learns options that are robust to model uncertainty. We utilize ROPI to learn robust options with the Robust Options Deep Q Network (RO-DQN) that solves multiple tasks and mitigates model misspecification due to model uncertainty. We present experimental results which suggest that policy iteration with linear features may have an inherent form of robustness when using coarse feature representations. In addition, we present experimental results which demonstrate that robustness helps policy iteration implemented on top of deep neural networks to generalize over a much broader range of dynamics than non-robust policy iteration.