BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.20.2//
METHOD:PUBLISH
X-WR-CALNAME;VALUE=TEXT:TSE
BEGIN:VTIMEZONE
TZID:Europe/Paris
BEGIN:STANDARD
DTSTART:20251026T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20260329T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.142007.field_date.0@www.tse-fr.eu
DTSTAMP:20260510T205912Z
CREATED:20260305T101001Z
DESCRIPTION:Franck Iutzeler (IMT - Université Paul Sabatier)\, “A study of 
 the long-run behavior of stochastic gradient descent via large deviations”
 \, MAD-Stat. Seminar\, Toulouse: TSE\, March 12\, 2026\, 11:00–12:15\, roo
 m Auditorium 3.\n\nWe examine the long-run distribution of stochastic grad
 ient descent (SGD) in general\, non-convex problems. Specifically\, we see
 k to understand which regions of the problem's state space are more likely
  to be visited by SGD\, and by how much. Using an approach based on the th
 eory of large deviations and randomly perturbed dynamical systems\, we \ns
 how that the long-run distribution of SGD resembles the Boltzmann-Gibbs di
 stribution of equilibrium thermodynamics with temperature equal to the met
 hod's step-size and energy levels determined by the problem's objective an
 d the statistics of the noise. \nJoint work w/ W. Azizian\, J. Malick\, P.
  Mertikopoulos
DTSTART;TZID=Europe/Paris:20260312T110000
DTEND;TZID=Europe/Paris:20260312T121500
LAST-MODIFIED:20260306T011001Z
LOCATION:Toulouse: TSE\, March 12\, 2026\, 11:00–12:15\, room Auditorium 3
SUMMARY:MAD-Stat. Seminar
URL;TYPE=URI:https://www.tse-fr.eu/seminars/2026-study-long-run-behavior-st
 ochastic-gradient-descent-large-deviations
END:VEVENT
END:VCALENDAR