BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:8940@i2m.univ-amu.fr
DTSTART;TZID=Europe/Paris:20251209T143000
DTEND;TZID=Europe/Paris:20251209T153000
DTSTAMP:20251205T162344Z
URL:https://www.i2m.univ-amu.fr/evenements/mean-field-analysis-of-the-trai
 ning-dynamics-of-two-layer-neural-networks-2/
SUMMARY:Arnaud Descours (ISFA\, Lyon): Mean-field analysis of the training 
 dynamics of two-layer neural networks
DESCRIPTION:Arnaud Descours: Training neural networks via stochastic gradie
 nt descent (SGD)\namounts to solving a complex\, non-convex optimization p
 roblem.\nInspired by statistical mechanics\, the mean-field approach provi
 des\na macroscopic description of the training dynamics\,\nwhich can be fo
 rmulated as a convex optimization problem.\nIn this talk\, I will explain 
 how to rigorously derive\nthe macroscopic (mean-field) description\nfrom t
 he microscopic dynamics given by the SGD updates.\nMore precisely\, I will
  establish the mean-field limit\n(a law of large numbers) and study the fl
 uctuations\naround this limit (a central limit theorem).\nIf time allows\,
  I will present similar results\nin the Bayesian framework.
CATEGORIES:Séminaire,ALEA,Probabilités
LOCATION:I2M Saint-Charles - Salle de séminaire\, Université Aix-Marseill
 e\, Campus Saint-Charles\, 3 Place Victor Hugo\, Marseille\, 13003\, Franc
 e
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=Université Aix-Marseille\,
  Campus Saint-Charles\, 3 Place Victor Hugo\, Marseille\, 13003\, France;X
 -APPLE-RADIUS=100;X-TITLE=I2M Saint-Charles - Salle de séminaire:geo:0,0
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20251026T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR