BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Drupal iCal API//EN
X-WR-CALNAME:Events items teaser
X-WR-TIMEZONE:America/Toronto
BEGIN:VTIMEZONE
TZID:America/Toronto
X-LIC-LOCATION:America/Toronto
BEGIN:DAYLIGHT
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
DTSTART:20230312T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20231105T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69d0d14fa7dfc
DTSTART;TZID=America/Toronto:20240202T120000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20240202T130000
URL:https://uwaterloo.ca/combinatorics-and-optimization/events/co-reading-g
 roup-rian-neogi-2
SUMMARY:C&O Reading Group - Rian Neogi
CLASS:PUBLIC
DESCRIPTION:TITLE: Regret bounds for FTRL and Mirror Descent\n\nSPEAKER:\n 
 Rian Neogi\n\nAFFILIATION:\n University of Waterloo\n\nLOCATION:\n MC 6029
 \n\nABSTRACT: In the previous talk\, I introduced FTRL and Mirror Descent\
 nand showed how they generalize two well-known algorithms of\nMultiplicati
 ve Weights and Gradient Descent. In this talk\, I will show\nthat FTRL and
  Mirror Descent are in fact equivalent in the sense that\nthey produce the
  same sequence of predictions. Moreover\, I will go\nover some regret boun
 ds for these algorithms\, that will generalize the\nregret bounds we get f
 or multiplicative weights and gradient descent.
DTSTAMP:20260404T085231Z
END:VEVENT
END:VCALENDAR