Workshop on Statistical Deep Learning

When: 24-25 October 2022

Where: Sydney Business School Level 10, 1 Macquarie Place, Sydney

About: This workshop aims to bring together researchers in the region that are interested in topics at the interface of deep learning and statistics. It will be based around a set of talks, and there will be time for discussion and networking between the talks. The workshop, which is an in-person event, is 1.5 days long, and will finish at noon on the 25th October. Refreshments and lunch on the 24th October are provided. Space is limited; please let us know at if you register, plans change, and you need to cancel registration.

Registration: The workshop is now very close to, or at, capacity, so the registration page has been closed. If you’d still like to attend, please send me an e-mail directly ( to check availability. (The workshop is free of charge).

Workshop Programme

Monday 24th October

08:30: Coffee

09:00: Maurizio Filippone (EURECOM): Functional priors for Bayesian deep learning
10:00: Michael Smith (UMelb): Marginally Calibrated Deep Distributional Regression

10:35: Coffee Break

11:10: Laura Cartwright (UOW): Emulation of Lagrangian particle dispersion model sensitivities using a variational autoencoder
11:45: Edwin Bonilla (CSIRO): Scalable inference in Gaussian process and deep Gaussian process models

12:20: Lunch

13:30: Laurent Jospin (UWA): Hands-on Bayesian neural networks - A tutorial for deep learning users
14:05: Luca Maestrini (ANU): Streamlined variational approximations for sparse linear mixed model selection

14:40: Coffee Break

15:15: Laurence Davies (QUT): Flow-enhanced transdimensional proposals
15:50: Quan Vu (UOW): Warped gradient-enhanced Gaussian process surrogate models
16:25: Close

18:00: Dinner (Chat Thai)

Tuesday 25th October

08:30: Coffee

09:00: Minh-Ngoc Tran (USyd): Bayesian deep net GLM and GLMM
09:35: Daniel Schmidt (Monash): Log-scale prior distributions for adaptive shrinkage

10:10: Coffee Break

10:40: Clara Grazian (USyd): Assessing the invertibility of deep biometric representations
11:15: Andrew Zammit Mangion (UOW): Deep neural Bayes estimators for spatial models
11:50: Close


This event is supported by the Australian Research Council (ARC DE180100203).

Image credit: Pratyeka. The header image is licensed under the Creative Commons Attribution-Share Alike 4.0 International license.