Learning the Linear Quadratic Regulator from Nonlinear Observations

Abstract

We introduce a new problem setting for continuous control called the LQR with Rich Observations, or RichLQR. In our setting, the environment is summarized by a low-dimensional continuous latent state with linear dynamics and quadratic costs, but the agent operates on high-dimensional, nonlinear observations such as images from a camera. To enable sample-efficient learning, we assume that the learner has access to a class of decoder functions (e.g., neural networks) that is flexible enough to capture the mapping from observations to latent states. We introduce a new algorithm, RichID, which learns a near-optimal policy for the RichLQR with sample complexity scaling only with the dimension of the latent state space and the capacity of the decoder function class. RichID is oracle-efficient and accesses the decoder class only through calls to a least-squares regression oracle. Our results constitute the first provable sample complexity guarantee for continuous control with an unknown nonlinearity in the system model and general function approximation.

arXiv commentTo appear at NeurIPS 2020

References in corpus (10)

• Learning Latent Dynamics for Planning from Pixels
• Empirical Bernstein Bounds and Sample Variance Penalization
• Contextual Decision Processes with Low Bellman Rank are PAC-Learnable
• Improper Learning for Non-Stochastic Control
• Logarithmic Regret Bound in Partially Observable Linear Dynamical Systems
• Provably efficient RL with Rich Observations via Latent State Decoding
• Robust Guarantees for Perception-Based Control
• Non-asymptotic and Accurate Learning of Nonlinear Dynamical Systems
• Predictive Coding for Locally-Linear Control
• Extracting Latent State Representations with Linear Dynamics from Rich Observations