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:20231029T030000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RDATE:20241027T030000
TZNAME:CET
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20240331T020000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.132303.field_date.0@www.tse-fr.eu
DTSTAMP:20260718T052828Z
CREATED:20230720T161001Z
DESCRIPTION:Sebastian Engelke (Université de Genève)\, “Machine learning be
 yond the data range: an extreme value perspective”\, MAD-Stat. Seminar\, T
 oulouse: TSE\, May 2\, 2024\, 11:00–12:15\, room Auditorium  5.\n\nMachine
  learning methods perform well in prediction tasks within the range of the
  training data.\nThese methods typically break down when interest is in (1
 ) prediction in areas of the predictor\nspace with few or no training obse
 rvations\; or (2) prediction of quantiles of the response that go\nbeyond 
 the observed records.\n\nExtreme value theory provides the mathematical fo
 undation for extrapolation beyond the range of\nthe training data\, both i
 n the dimension of the predictor space and the response variable. In this
 \ntalk we present recent methodology that combines this extrapolation theo
 ry with flexible machine\nlearning methods to tackle the out-of-distributi
 on generalization problem (1) and the extreme\nquantile regression problem
  (2).\n\nWe show the practical importance of prediction beyond the trainin
 g observations in environmental\nand climate applications\, where domain s
 hifts in the predictor space occur naturally due to climate\nchange and ri
 sk assessment for extreme quantiles is required.
DTSTART;TZID=Europe/Paris:20240502T120000
DTEND;TZID=Europe/Paris:20240502T131500
LAST-MODIFIED:20240420T001001Z
LOCATION:Toulouse: TSE\, May 2\, 2024\, 11:00–12:15\, room Auditorium  5
SUMMARY:MAD-Stat. Seminar
URL;TYPE=URI:https://www.tse-fr.eu/seminars/2024-machine-learning-beyond-da
 ta-range-extreme-value-perspective
END:VEVENT
END:VCALENDAR