Wednesday, September 30, 2020 - 4:00pm to 4:30pm
Event Calendar Category
LIDS & Stats Tea
Zoom meeting id
916 6732 9861
Join Zoom meeting
While the identification of nonlinear dynamical systems is a fundamental building block of model-based reinforcement learning and feedback control, its sample complexity is only understood for systems that either have discrete states and actions or for systems that can be identified from data generated by i.i.d. random inputs. Nonetheless, many interesting dynamical systems have continuous states and actions and can only be identified through a judicious choice of inputs. Motivated by practical settings, we study a class of nonlinear dynamical systems whose state transitions depend linearly on a known feature embedding of state-action pairs. To estimate such systems infinite time identification methods must explore all directions in feature space. We propose an active learning approach that achieves this by repeating three steps: trajectory planning, trajectory tracking, and re-estimation of the system from all available data. We show that our method estimates nonlinear dynamical systems at a parametric rate, similar to the statistical rate of standard linear regression.
Horia is a postdoc in LIDS, co-hosted by Ali Jadbabaie, Devavrat Shah, and Suvrit Sra. He is broadly interested in the foundations of machine learning with a focus on characterizing the sample complexity of reinforcement learning. Horia received his PhD in Computer Science from UC Berkeley, under the guidance of Michael I. Jordan and Benjamin Recht. Before moving to Berkeley, Horia obtained a BA in Mathematics from Princeton University.