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:20210314T070000
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:20201101T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:69d2b64437922
DTSTART;TZID=America/Toronto:20210625T153000
SEQUENCE:0
TRANSP:TRANSPARENT
DTEND;TZID=America/Toronto:20210625T153000
URL:https://uwaterloo.ca/combinatorics-and-optimization/events/tutte-colloq
 uium-dmitriy-drusvyatskiy
SUMMARY:Tutte Colloquium - Dmitriy Drusvyatskiy
CLASS:PUBLIC
DESCRIPTION:TITLE: From low probability to high confidence in stoch
 astic convex optimization\n\nSpeaker:\n Dmitriy Drusvyatskiy\n\nAfflilia
 tion:\n University of Washington\n\nZoom:\n Contact Emma Watson\n\nABSTRA
 CT:\n\nStandard results in stochastic convex optimization bound the number
  of\nsamples that an algorithm needs to generate a point with small\nfunct
 ion value in expectation. More nuanced high probability\nguarantees are ra
 re\, and typically either rely on “light-tail”\nnoise assumptions or e
 xhibit worse sample complexity. In this work\, we\nshow that a wide class 
 of stochastic optimization algorithms can be\naugmented with high confiden
 ce bounds at an overhead cost that is only\nlogarithmic in the confidence 
 level and polylogarithmic in the\ncondition number.
DTSTAMP:20260405T192140Z
END:VEVENT
END:VCALENDAR