BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Francesco Locatello: Learning disentangled representations
DTSTART:20191023T141500
DTEND:20191023T150000
DTSTAMP:20260604T032156Z
UID:236cb01e99141892205559ff95f5c8537ba83577c0e2c53552d88d32
CATEGORIES:Conferences - Seminars
DESCRIPTION:Francesco Locatello\nAbstract: \nThe key idea behind the unsu
 pervised learning of disentangled representations is that real-world data 
 is generated by few explanatory factors of variation (e.g. content + posit
 ion of objects in an image) which can be recovered by unsupervised learnin
 g algorithms. \nIn this talk\, I will discuss the recent progress in the 
 field and challenge some common assumptions. I will discuss the role of in
 ductive biases in the theoretical impossibility of unsupervised learning o
 f disentangled representations and provide a sober look at the performance
 s of state-of-the-art approaches. I will further address the question of h
 ow to go beyond purely unsupervised disentanglement in both theory and pra
 ctice and discuss applications to fairness and visual abstract reasoning.\
 n\nBio:\nFrancesco Locatello is a Doctoral Fellow at the Max Planck ETH Ce
 nter for Learning Systems supervised by Gunnar Rätsch and Bernhard Schöl
 kopf. He is also working as a research consultant at Google Brain in Züri
 ch. He holds a Google PhD Fellowship in Machine Learning and his work on d
 isentangled representations received a Best Paper Award at ICML 2019.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR