Generative Models and Uncertainty Quantification

Copenhagen, October 12 and 13, 2021

Generative models and uncertainty quantification lie at the heart of Bayesian modelling and inference. At this small meeting, we discuss recent developments within the field. The meeting is deliberately kept small in order to ensure that discussion remains honest, lively and interesting. Attendance is, thus, mostly by invitation, but one can apply to join (see below).

Confirmed Speakers

Eric Nalisnick
Assistant Professor, University of Amsterdam.

Yarin Gal
Associate Professor, University of Oxford.

Jakub Tomczak
Assistant professor, Vrije Universiteit Amsterdam.

Clare Lyle
PhD student, University of Oxford.

Marine Le Morvan
Postdoctoral researcher, INRIA Saclay.

Jan-Willem van de Meent
Associate Professor & Co-director AMLab, University of Amsterdam.

Mark van der Wilk
Assistant Professor, Imperial College London.

Matthias Bauer
Research Scientist, Deepmind London.

Nicolas Durrande
Director of Research at Secondmind.

Markus Heinonen
Research Fellow, Aalto University and Finnish Center of AI.

Aude Sportisse
Postdoctoral researcher, INRIA Sophia-Antipolis.

Program

Tuesday (Oct 12)		Wednesday (Oct 13)
Session 1 Bayesian deep learning Chair: Wouter Boomsma		Session 4 Generative models Chair: Søren Hauberg
9:25-9:30	Opening remarks Jes Frellsen
9:30-10:30	Matthias Bauer Revisiting the Laplace Approximation for Model Selection in BNNs	09:00-10:00	Mark van der Wilk Inductive Biases, Input Densities, and Predictive Uncertainty
10:30-10:50	Coffee	10:00-10:20	Coffee
10:50-11:30	Eric Nalisnick Predictive Complexity Priors	10:20-11:00	Jan-Willem van de Meent Compositional Inference in Probabilistic Programs
11:30-12:30	Clare Lyle Bayesian Model Selection and Generalization in Deep Learning	11:00-12:00	Jakub Tomczak Is the Likelihood-based Deep Generative Modeling appropriate for Representation Learning?
12:30-13:30	Lunch	12:00-13:00	Lunch
Session 2 Missing data Chair: Pierre-Alexandre Mattei		Session 5 Nonparametrics Chair: Ole Winther
13:30-14:30	Marine Le Morvan What's a good imputation to predict with missing values?	13:00-14:00	Yarin Gal Non-Bayesian Nonparametrics
14:30-15:10	Yarin Gal Surprice talk...	14:00-14:40	Nicolas Durrande Designing inducing variables that connect sparse Gaussian process models and Neural Networks
15:10-15:30	Coffee	14:40-15:00	Coffee
Session 3 Dynamics Chair: Jes Frellsen		15:00-15:10	Concluding remark
15:30-16:10	Markus Heinonen Learning continuous-time dynamics
16:10-17:30	Poster Session
18:00-	Conference Dinner at restaurant Høst (directions)

Registration

If you would like to attend this workshop but have yet to receive an invitation, then please write an email to Søren Hauberg and include a link to your website or Google scholar profile. Note that seating is limited, so we cannot guarantee a ticket.

Venue

The workshop takes place at:

Ceremonial Hall (Festsalen), University of Copenhagen
Frue Plads 4
1168 København K, Denmark

Pier47

Organizers

The conference is jointly organized by:

This workshop is funded by the Novo Nordisk Foundation with the workshop grant NNF19OC0059371 and through the Center for Basic Machine Learning Research in Life Science (NNF20OC0062606).