BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/1743
DTSTAMP:20260707T080944Z
SUMMARY:Model-Free Reinforcement Learning as Constrained Online Convex Opti
 mization
DESCRIPTION:Speaker: Aakash Ghosh (TIFR)\n\nAbstract: \n\nClassical model-f
 ree control treats the optimal action-value function Q* as the fixed point
  of the Bellman optimality operator and chases it by stochastic approximat
 ion. This talk develops a different view: Q* is the unique minimizer of an
  exact\, unregularized linear program whose constraints are the Bellman in
 equalities\, and a single agent–environment trajectory is a stream of sa
 mpled rows of that program. Learning Q* then becomes an instance of Constr
 ained Online Convex Optimization (COCO)\, which we solve with a drift-plus
 -penalty algorithm that maintains one virtual queue of accumulated Bellman
  violations per state–action pair.\nWe show the time-averaged table conv
 erges to Q* from a single Markovian trajectory. The reduction is modular a
 nd any optimizer meeting a regret/violation/movement interface plugs in.\n
URL:https://www.tcs.tifr.res.in/web/events/1743
DTSTART;TZID=Asia/Kolkata:20260708T163000
DTEND;TZID=Asia/Kolkata:20260708T173000
LOCATION:A-201 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR