BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Self-Supervised Representation Learning  - Javier Antoran (Univers
 ity of Cambridge)
DTSTART:20200226T110000Z
DTEND:20200226T123000Z
UID:TALK140320@talks.cam.ac.uk
CONTACT:75379
DESCRIPTION:Self-Supervised Learning (SSL)\, combined with deep neural net
 work models\, has seen great success recently. Representations learnt with
  SSL have been used to obtain state-of-the-art results on imagenet as well
  as text and speech classification tasks. We begin by introducing SSL\, fo
 cusing on two of the most successful approaches: contrastive predictive co
 ding (CPC) and deep InfoMax (DIM). We provide information theoretic interp
 retations of these approaches\, and discuss the role of mutual information
  as a principal motivator for this framework.\nNext\, we consider the role
  of identifiability in representation learning with generative models. We 
 discuss recent results demonstrating that seemingly heuristic SSL approach
 es can be used to guarantee identifiability in nonlinear-ICA. These result
 s imply an interesting link between identifiable generative models and SSL
 \, potentially providing a principled foundation for SSL approaches.
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR