BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Learning pipelines bridging reinforcement learning and adaptive co
 ntrol
DTSTART:20261028T151500
DTEND:20261028T161500
DTSTAMP:20261019T155020Z
UID:e00d8191de3cbc7504a3d8da1171235eae84c54ac4861ad797df81b2
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr Feiran Zhao\, postdoctoral researcher at Automatic Control 
 Lab in ETH Zurich\nAbstract: \nThe adjacent fields of reinforcement learn
 ing (RL) and adaptive control share the same objectives\, yet they are sep
 arated by a wide cultural gap. In this presentation\, I attempt to bridge 
 this gap for the linear quadratic regulator (LQR) problem\, which serves a
 s a cornerstone and the benchmark for both fields. I begin by discussing d
 ifferent learning pipelines\, including direct and indirect (model-based) 
 approaches\, as well as episodic and online (adaptive) approaches. Despite
  the extensive literature spanning several decades\, numerous problems rem
 ain unsolved. For instance\, RL methods are seldom concerned with closed-l
 oop stability certificates or efficient implementations\, while the adapti
 ve control community has dedicated minimal effort to optimality. We addres
 s the data-driven LQR problem in an adaptive setting\, which entails onlin
 e recursive algorithms and closed-loop data\, and we seek both algorithmic
  as well as closed-loop certificates. Our approach encompasses different v
 ariations of policy gradient methods going also beyond mere LQR formulatio
 ns\, employs a novel covariance parameterization\, and leverages different
  kind of regularizations. Finally\, all our theoretical results are valida
 ted through simulations and experiments in diverse domains\, demonstrating
  the computational and sample efficiency of our method.\n\nBiography: \nF
 eiran Zhao is currently a postdoctoral researcher at Automatic Control Lab
  in ETH Zurich\, working with Prof. Florian Dörfler. Before that\, he re
 ceived the Ph.D. Degree in Control Science and Engineering from the Tsingh
 ua\nUniversity\, China\, in 2024. His research interests include reinforce
 ment learning\, control theory\, optimization\, their intersections\, and 
 applications on energy\, robotics\, and mobility systems.\n 
LOCATION:ME C2 405 https://plan.epfl.ch/?room==ME%20C2%20405
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR