BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Recent advances in the theory and applications of VAEs - Gergely F
 lamich and  Alexandru Cioba
DTSTART:20200617T100000Z
DTEND:20200617T113000Z
UID:TALK149596@talks.cam.ac.uk
CONTACT:75379
DESCRIPTION:Since their development by Kingma et.al.[1] the VAE has proven
  to be a flexible and powerful framework for latent variable modelling. Si
 tting at the crossroads of classical statistical learning and deep learnin
 g\, it has made an impact in many problems\, from classical tasks in compu
 ter vision such as image generation\, compression\, super-resolution\, inp
 ainting\, 3D object synthesis to other tasks such as semi-supervised learn
 ing\, natural language modelling\, sentence interpolation\, and even pract
 ical applications to medical imaging. \n\nWe will start with a minimal int
 roduction to latent variable models and the structure and definition of a 
 VAE\, before presenting some examples. We will then discuss some of the to
 pics of interest in the research surrounding VAEs\, looking to various tec
 hniques meant to reduce the inference gap\, as well as studying the useful
 ness of the VAE for learning latent representations for downstream tasks. 
 If time allows\, we will present some more off-beat results.\n\n[1]: Auto-
 encoding variational Bayes\; D. Kingma\, M. Welling: https://arxiv.org/abs
 /1312.6114
LOCATION:Venue to be confirmed
END:VEVENT
END:VCALENDAR