BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250916T001850EDT-4146ZNbxnI@132.216.98.100
DTSTAMP:20250916T041850Z
DESCRIPTION:Title: Robust Risk-Aware Reinforcement Learning.\n\nAbstract: W
 e present a reinforcement learning (RL) approach for robust optimisation o
 f risk-aware performance criteria. To allow agents to express a wide varie
 ty of risk-reward profiles\, we assess the value of a policy using rank de
 pendent expected utility (RDEU). RDEU allows the agent to seek gains\, whi
 le simultaneously protecting themselves against downside risk. To robustif
 y optimal policies against model uncertainty\, we assess a policy not by i
 ts distribution\, but rather\, by the worst possible distribution that lie
 s within a Wasserstein ball around it. Thus\, our problem formulation may 
 be viewed as an actor/agent choosing a policy (the outer problem)\, and th
 e adversary then acting to worsen the performance of that strategy (the in
 ner problem). We develop explicit policy gradient formulae for the inner a
 nd outer problems\, and show its efficacy on three prototypical financial 
 problems: robust portfolio allocation\, optimising a benchmark\, and stati
 stical arbitrage. [ This is joint work with Silvana Pesenti\, Ye Shegn Wan
 g\, and Hariom Tatsat ]\n\n \n\n \n\nSeminar Quantact\n	Zoom meeting\n
DTSTART:20211112T190000Z
DTEND:20211112T200000Z
SUMMARY:Sebastian Jaimungal (University of Toronto)
URL:/mathstat/channels/event/sebastian-jaimungal-unive
 rsity-toronto-334656
END:VEVENT
END:VCALENDAR