5th Symposium on
Advances in Approximate Bayesian Inference

Sunday July 23rd, 2023 at the Ala Moana Hotel, 410 Atkinson Drive in Honolulu, Hawaii 96814 United States Co-located with ICML


Schedule

Authors presenting posters will be given their poster location at registration. Due to the large number of posters to be presented, we will be having two poster sessions. Authors presenting in the first poster sessions should arrive 10 minutes before its start to put up their posters. Authors presenting in the second poster session should arrive 10 minutes before the end of the lunch break to put up their posters.

Our talks will be divided into 3 thematic sessions. Invited talks will last 40 minutes (30 minutes talk + 10 minutes questions). Contributed talks will last 15 minutes (10 minutes talk + 5 minutes questions).


8:00 - 9:00 Registration + Breakfast
8:30 - 10:00 Poster Session 1 (OpenReview paper IDs: 22-41 Regular track and 10-19 Fast track )
Session 1: Generative Models
Chair: Javier Antorán
10:00 - 10:10 Opening Remarks
10:10 - 10:50 Invited Approximate Bayesian Inference, Algorithms, and Diffusion Models; Is Our Gig Up?
Frank Wood
10:50 - 11:05 Contributed How to Train Your FALCON: Learning Log-Concave Densities with Energy-Based Neural Networks
Alexander Lin, Demba E. Ba
11:05 - 11:45 Invited Diffusion Schrödinger Bridge Matching
Valentin de Bortoli
11:45 - 12:00 Contributed Neural Adaptive Smoothing via Twisting
Michael Y. Li, Dieterich Lawson, Scott Linderman
12:00 - 13:00 Lunch Break (On your own)
Session 2: VI and MCMC
Chair: Ruqi Zhang
13:00 - 13:40 Invited Gaussian-Bernoulli RBMs Without Tears
Renjie Liao
13:40 - 13:55 Contributed Automatically Marginalized MCMC in Probabilistic Programming
Jinlin Lai, Javier Burroni, Hui Guan, Daniel Sheldon
13:55 - 14:35 Invited Probabilistic inference when the model is wrong
Diana Cai
14:35 - 14:50 Contributed Variational Prediction
Alexander A Alemi, Ben Poole
14:50 - 16:00 Poster Session 2 (OpenReview paper IDs: 1-21 Regular track and 1-9 Fast track )
+ Coffee Break
Session 3: Bayesian Neural Networks
Chair: Vincent Fortuin
16:00 - 16:40 Invited Bayesian Autoencoders
Maurizio Fillipone
16:40 - 16:55 Contributed Online Laplace Model Selection Revisited
Jihao Andreas Lin, Javier Antorán, José Miguel Hernández-Lobato
16:55 - 17:55 Panel Discussion: Approximate inference in the age of generative AI
Moderator: Theofanis Karaletsos
17:55 - 18:00 Closing Remarks

Questions? Contact us at symposium.approximateinference@gmail.com.