BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20251007T023606EDT-4881v5L1kk@132.216.98.100
DTSTAMP:20251007T063606Z
DESCRIPTION:Progress in theoretical understanding of deep learning.\n\nDeep
  learning has arisen around 2006 as a renewal of neural networks research 
 allowing such models to have more layers. Theoretical investigations have 
 shown that functions obtained as deep compositions of simpler functions (w
 hich includes both deep and recurrent nets) can express highly varying fun
 ctions (with many ups and downs and different input regions that can be di
 stinguished) much more efficiently (with fewer parameters) than otherwise\
 , under a prior which seems to work well for artificial intelligence tasks
 . Empirical work in a variety of applications has demonstrated that\, when
  well trained\, such deep architectures can be highly successful\, remarka
 bly breaking through previous state-of-the-art in many areas\, including s
 peech recognition\, object recognition\, language models\, machine transla
 tion and transfer learning. Although neural networks have long been consid
 ered lacking in theory and much remains to be done\, theoretical advances 
 have been made and will be discussed\, to support distributed representati
 ons\, depth of representation\, the non-convexity of the training objectiv
 e\, and the probabilistic interpretation of learning algorithms (especiall
 y of the auto-encoder type\, which were lacking one). The talk will focus 
 on the intuitions behind these theoretical results.\n
DTSTART:20161118T203000Z
DTEND:20161118T213000Z
LOCATION:room 1205\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Yoshua Bengio\, Université de Montréal
URL:/mathstat/channels/event/yoshua-bengio-universite-
 de-montreal-264235
END:VEVENT
END:VCALENDAR