BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Variational Bayes as Surrogate Regression - Will Tebbutt (Universi
 ty of Cambridge)
DTSTART:20210210T110000Z
DTEND:20210210T123000Z
UID:TALK156724@talks.cam.ac.uk
CONTACT:Elre Oldewage
DESCRIPTION:Variational Bayes is a useful approximate inference framework 
 in which an intractable posterior distribution is approximated by simpler 
 tractable one. \nThe extent to which this is useful (usually) depends on h
 ow closely this approximation matches reality\, and how quickly it can be 
 obtained. We'll present lines of work that utilise the posteriors of tract
 able models as this approximation\, and the interesting inference algorith
 ms that arise in this setting. \n\nAlthough we'll cover all of these in th
 e presentation\, it will be helpful to have some familiarity with the basi
 cs of variational Bayes (e.g. what the ELBO is)\, variational autoencoders
  and the idea of amortised inference\, exponential families\, and Gaussian
  processes. A basic understanding of natural gradients would also be helpf
 ul\, but is\nnot essential. \n \nIf you have the time\, please read this: 
 \nOpper\, Manfred\, and Cédric Archambeau. "The variational Gaussian\napp
 roximation revisited." Neural computation 21.3 (2009): 786-792.\n\nExtra r
 eading if you have time on your hands: \n# Bui\, Thang D.\, et al. "Partit
 ioned variational inference: A unified framework encompassing federated an
 d continual learning." arXiv preprint arXiv:1811.11206 (2018).\n# Ashman\,
  Matthew\, et al. "Sparse Gaussian Process Variational Autoencoders." arXi
 v preprint arXiv:2010.10177 (2020). Khan\, Mohammad Emtiyaz\, and Didrik N
 ielsen. "Fast yet simple natural-gradient descent for variational inferenc
 e in complex models." 2018 International Symposium on Information Theory a
 nd Its Applications (ISITA). IEEE\, 2018.\n# Chang\, Paul E.\, et al. "Fas
 t variational learning in state-space Gaussian process models." 2020 IEEE 
 30th International Workshop on Machine Learning for Signal Processing (MLS
 P). IEEE\, 2020. \n# Johnson\, Matthew James\, et al. "Composing graphical
  models with neural networks for structured representations and fast infer
 ence." Proceedings of the 30th International Conference on Neural Informat
 ion Processing Systems. 2016.
LOCATION:https://eng-cam.zoom.us/j/86068703738?pwd=YnFleXFQOE1qR1h6Vmtwbno
 0LzFHdz09
END:VEVENT
END:VCALENDAR