CAM Colloquium: Sarah Dean (Computer Science, Cornell) - Data-driven Control and Decision-making in Feedback Systems

Description

Abstract: 

Machine learning techniques have been successful for processing complex information, and thus they have the potential to play an important role in data-driven decision-making and control. However, ensuring the reliability of these methods in feedback systems remains a challenge, since classic statistical and algorithmic guarantees do not always hold.

In this talk, I will discuss dynamical settings relevant to robotics and recommendation systems. We will primarily consider the case that the dynamics of the system are unknown, so control or decision policies must be based on observed data. Starting with linear optimal control, I will discuss a simple approach with finite-sample performance and safety guarantees. Moving onto nonlinear dynamics, I will show how to robustify controllers to provide data-dependent safety and stability guarantees. Finally, I will discuss ongoing work on dynamics that may arise in the setting of recommendation systems.

Bio: Sarah is an incoming Assistant Professor in the Computer Science Department at Cornell. She recently completed a PhD in EECS from UC Berkeley and was a postdoc at the University of Washington. Sarah is interested in the interplay between optimization, machine learning, and dynamics, and her research focuses on understanding the fundamentals of data-driven control and decision-making.