BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning in Continuous-Time Linear-Quadratic Games with Heterogene
 ous Players - Philipp Plank (Imperial College London)
DTSTART:20251112T113000Z
DTEND:20251112T121000Z
UID:TALK238495@talks.cam.ac.uk
DESCRIPTION:Multi-agent reinforcement learning\, despite its popularity an
 d empirical success\, faces significant scalability challenges in large-po
 pulation dynamic games\, especially with heterogeneous players. This talk 
 will use fundamental linear-quadratic games as an example and present rece
 nt frameworks that provide principled designs for efficient and scalable l
 earning algorithms in multi-agent systems with heterogeneous players.\nIn 
 the first part\, we will introduce the Graphon Mean Field Game approach an
 d present provably convergent policy gradient algorithms for large-populat
 ion games in which agents interact weakly through a symmetric graph. The s
 econd part of the talk will focus on the Alpha-Potential Game framework\, 
 which enables the development of efficient learning algorithms for asymmet
 ric network games that go beyond mean-field approximations.\nThis talk is 
 based on joint work with Yufei Zhang.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR