BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:www.tcs.tifr.res.in/event/1775
DTSTAMP:20260911T050427Z
SUMMARY:When in Doubt\, Abstain: Fast Rates for Online Prediction
DESCRIPTION:Speaker: Aakash Ghosh (TIFR)\n\nAbstract: \nSince this talk fal
 ls on 9/11\, we will be careful about confident predictions: sometimes the
  right action is to abstain.\nWe study online prediction with N experts an
 d binary outcomes. Without abstention\, the standard regret rate is Θ(√
 (T log N)). Neu and Zhivotovskiy show that allowing the learner to abstain
  at cost c < 1/2 changes the rate to Θ(min{log N / (1 − 2c)\, √(T log
  N)}).\nThus\, for fixed c < 1/2\, regret eventually stops growing with th
 e number of rounds.\nWe will derive the algorithm from exponential weights
 \, explain why its abstention probability tracks expert disagreement\, and
  prove the upper bound through a one-variable inequality. We will then dis
 cuss the matching lower bound\, varying abstention costs\, and extensions 
 to multiclass and infinite expert classes. \n
URL:https://www.tcs.tifr.res.in/web/events/1775
DTSTART;TZID=Asia/Kolkata:20260911T160000
DTEND;TZID=Asia/Kolkata:20260911T170000
LOCATION:A-201 (STCS Seminar Room)
END:VEVENT
END:VCALENDAR