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:6658@i2m.univ-amu.fr
DTSTART;TZID=Europe/Paris:20201019T140000
DTEND;TZID=Europe/Paris:20201019T160000
DTSTAMP:20241120T201929Z
URL:https://www.i2m.univ-amu.fr/evenements/mariia-vladimirova-prior-specif
 ication-for-bayesian-deep-learning-models/
SUMMARY: (...): Mariia Vladimirova : Prior specification for Bayesian deep 
 learning models
DESCRIPTION:: Title:\nPrior specification for Bayesian deep learning models
 \n\nAbstract:\nNeural networks (NNs)\, and their deep counterparts\, have 
 largely been used in many research areas such as image analysis\, signal p
 rocessing\, or reinforcement learning\, just to name a few. The impressive
  performance provided by such machine learning approaches has greatly moti
 vated research that aims at a better understanding the driving mechanisms 
 behind their effectiveness. In particular\, the study of the NNs distribut
 ional properties through Bayesian analysis has recently gained much attent
 ion.\n\nIn this seminar we firstly describe the necessary notations and st
 atistical background for Bayesian NNs. Then we consider its distributional
  properties and novel theoretical insight on distributions at the units le
 vel. Under the assumption of independent and normally distributed weights\
 , we establish that the induced prior distribution on the units before and
  after activation becomes increasingly heavy-tailed with the depth of the 
 layer. Lastly\, we discuss this property in terms of a regularizing mechan
 ism and corroborate it with experimental simulation results.\n\n&nbsp\;
CATEGORIES:Séminaire,Statistique
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:DAYLIGHT
DTSTART:20200329T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR