BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.4.0.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:130@isdm.umontpellier.fr
DTSTART;TZID=Europe/Paris:20250217T160000
DTEND;TZID=Europe/Paris:20250217T160000
DTSTAMP:20260602T130232Z
URL:https://isdm.umontpellier.fr/events/rethinking-early-stopping-refine-t
 hen-calibrate-2/
SUMMARY:Rethinking Early Stopping: Refine\, Then Calibrate
DESCRIPTION:Inria Montpellier\, St-Priest Campus\, Building 5\, Room 02/022
 \nMachine Learning in Montpellier\, Theory &amp\; Practice\n\nMachine lear
 ning classifiers often produce probabilistic predictions that are critical
  for accurate and interpretable decision-making in various domains. The qu
 ality of these predictions is generally evaluated with proper losses like 
 cross-entropy\, which decompose into two components: calibration error ass
 esses general under/overconfidence\, while refinement error measures the a
 bility to distinguish different classes.\nIn this paper\, we provide theor
 etical and empirical evidence that these two errors are not minimized simu
 ltaneously during training. Selecting the best training epoch based on val
 idation loss thus leads to a compromise point that is suboptimal for both 
 calibration error and\, mostimportantly\, refinement error. To address thi
 s\, we introduce a new metric for early stopping and hyperparameter tuning
  that makes it possible to minimize refinement error during training. The 
 calibration error is minimized after training\, using standard techniques.
  Our method integrates seamlessly with any architecture and consistently i
 mproves performance across diverse classification tasks.\n\nLabéliséHall
 esIA
ATTACH;FMTTYPE=image/jpeg:https://isdm.umontpellier.fr/wp-content/uploads/
 2025/02/ml-mpt-1.jpg
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20241027T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR