This site provides supplemental material to the paper Algorithmic Framework for Model-based Depp Reinforcement Learning with Theoretical Guarantees

Abstract

Model-based reinforcement learning (RL) is considered to be a promising approach to reduce the sample complexity that hinders model-free RL. However, the theoretical understanding of such methods has been rather limited. This paper introduces a novel algorithmic framework for designing and analyzing model-based RL algorithms with theoretical guarantees. We design a meta-algorithm with a theoretical guarantee of monotone improvement to a local maximum of the expected reward. The meta-algorithm iteratively builds a lower bound of the expected reward based on the estimated dynamical model and sample trajectories, and then maximizes the lower bound jointly over the policy and the model. The framework extends the optimism-in-face-of-uncertainty principle to non-linear dynamical models in a way that requires no explicit uncertainty quantification. Instantiating our framework with simplification gives a variant of model-based RL algorithms Stochastic Lower Bounds Optimization (SLBO). Experiments demonstrate that SLBO achieves state-of-the-art performance when only one million or fewer samples are permitted on a range of continuous control benchmark tasks.

Code will be made available upon approval.

Videos below show the real trajectories from the real environment (MuJoCo) and the imaginary trajectories/rollouts from the estimated models (neural nets) from the same starting position at various iterations. We found, somewhat surprisingly, the estimated model can predict the future states relatively accurately (when the actions are from the learned policy.)