Workshop
Your Model is Wrong: Robustness and misspecification in probabilistic modeling
Diana Cai · Sameer Deshpande · Michael Hughes · Tamara Broderick · Trevor Campbell · Nick Foti · Barbara Engelhardt · Sinead Williamson
Tue 14 Dec, 4:55 a.m. PST
Probabilistic modeling is a foundation of modern data analysis -- due in part to the flexibility and interpretability of these methods -- and has been applied to numerous application domains, such as the biological sciences, social and political sciences, engineering, and health care. However, any probabilistic model relies on assumptions that are necessarily a simplification of complex real-life processes; thus, any such model is inevitably misspecified in practice. In addition, as data set sizes grow and probabilistic models become more complex, applying a probabilistic modeling analysis often relies on algorithmic approximations, such as approximate Bayesian inference, numerical approximations, or data summarization methods. Thus in many cases, approximations used for efficient computation lead to fitting a misspecified model by design (e.g., variational inference). Importantly, in some cases, this misspecification leads to useful model inferences, but in others it may lead to misleading and potentially harmful inferences that may then be used for important downstream tasks for, e.g., making scientific inferences or policy decisions.
The goal of the workshop is to bring together researchers focused on methods, applications, and theory to outline some of the core problems in specifying and applying probabilistic models in modern data contexts along with current state-of-the-art solutions. Participants will leave the workshop with (i) exposure to recent advances in the field, (ii) an idea of the current major challenges in the field, and (iii) an introduction to methods meeting these challenges. These goals will be accomplished through a series of invited and contributed talks, poster spotlights, poster sessions, as well as ample time for discussion and live Q&A.
Schedule
Tue 4:55 a.m. - 5:00 a.m.
|
Welcome remarks
(
Talk
)
>
SlidesLive Video |
Diana Cai 🔗 |
Tue 5:00 a.m. - 5:30 a.m.
|
How to train your model when it's wrong: Bayesian nonparametric learning in M-open
(
Invited Talk
)
>
SlidesLive Video |
Chris C Holmes 🔗 |
Tue 5:30 a.m. - 5:35 a.m.
|
Invite Talk 1 Q&A ( Q&A ) > link | Chris C Holmes 🔗 |
Tue 5:35 a.m. - 6:05 a.m.
|
BayesCG: A probabilistic numeric linear solver
(
Invited Talk
)
>
SlidesLive Video |
Ilse Ipsen 🔗 |
Tue 6:05 a.m. - 6:10 a.m.
|
Invited Talk 2 Q&A
(
Q&A
)
>
|
Ilse Ipsen 🔗 |
Tue 6:10 a.m. - 6:45 a.m.
|
Individual discussions in Gathertown ( Gathertown discussion ) > link | 🔗 |
Tue 6:45 a.m. - 7:00 a.m.
|
Bayesian Calibration of imperfect computer models using Physics-informed priors
(
Contributed Talk
)
>
SlidesLive Video |
Michail Spitieris 🔗 |
Tue 7:00 a.m. - 7:15 a.m.
|
Invariant Priors for Bayesian Quadrature
(
Contributed Talk
)
>
SlidesLive Video |
Masha Naslidnyk 🔗 |
Tue 7:15 a.m. - 8:30 a.m.
|
Poster Session I in Gathertown ( Poster session ) > link | 🔗 |
Tue 8:30 a.m. - 9:35 a.m.
|
Research panel
(
Discussion panel
)
>
SlidesLive Video |
David Dunson · Marta Kwiatkowska · Steven MacEachern · Jeffrey Miller · Briana Joy Stephenson · Anirban Bhattacharya 🔗 |
Tue 9:35 a.m. - 10:30 a.m.
|
Individual discussions in Gathertown ( Gathertown discussion ) > link | 🔗 |
Tue 10:30 a.m. - 10:45 a.m.
|
Uncertainty estimation under model misspecification in neural network regression
(
Contributed Talk
)
>
SlidesLive Video |
Maria Cervera 🔗 |
Tue 10:45 a.m. - 11:00 a.m.
|
Your Bandit Model is Not Perfect: Introducing Robustness to Restless Bandits Enabled by Deep Reinforcement Learning
(
Contributed Talk
)
>
SlidesLive Video |
Jackson Killian 🔗 |
Tue 11:00 a.m. - 11:30 a.m.
|
Bayesian Model Averaging is not Model Combination: A PAC-Bayesian Analysis of Deep Ensembles
(
Invited Talk
)
>
SlidesLive Video |
Andres Masegosa 🔗 |
Tue 11:00 a.m. - 11:05 a.m.
|
Invited Talk 3 Q&A
(
Q&A
)
>
|
Andres Masegosa 🔗 |
Tue 11:35 a.m. - 12:00 p.m.
|
Individual discussions in Gathertown ( Gathertown discussion ) > link | 🔗 |
Tue 12:00 p.m. - 12:15 p.m.
|
PAC^m-Bayes: Narrowing the Empirical Risk Gap in the Misspecified Bayesian Regime
(
Contributed Talk
)
>
SlidesLive Video |
Alexander Alemi 🔗 |
Tue 12:15 p.m. - 12:30 p.m.
|
Bayesian Data Selection
(
Contributed Talk
)
>
SlidesLive Video |
Eli N Weinstein 🔗 |
Tue 12:30 p.m. - 1:00 p.m.
|
Statistically Robust Inference with Stochastic Gradient Algorithms
(
Invited Talk
)
>
SlidesLive Video |
Jonathan Huggins 🔗 |
Tue 1:00 p.m. - 1:05 p.m.
|
Invited Talk 4 Q&A
(
Q&A
)
>
|
Jonathan Huggins 🔗 |
Tue 1:05 p.m. - 1:30 p.m.
|
Individual discussions in Gathertown ( Gathertown discussion ) > link | 🔗 |
Tue 1:30 p.m. - 2:00 p.m.
|
Your Model is Wrong (but Might Still Be Useful)
(
Invited Talk
)
>
SlidesLive Video |
Lester Mackey 🔗 |
Tue 2:00 p.m. - 2:05 p.m.
|
Invited Talk 5 Q&A
(
Q&A
)
>
|
Lester Mackey 🔗 |
Tue 2:05 p.m. - 2:35 p.m.
|
Statistical and Computational Tradeoffs in Variational Bayes
(
Invited Talk
)
>
SlidesLive Video |
Yixin Wang 🔗 |
Tue 2:35 p.m. - 2:40 p.m.
|
Invited Talk 6 Q&A
(
Q&A
)
>
|
Yixin Wang 🔗 |
Tue 3:15 p.m. - 4:30 p.m.
|
Poster session II in Gathertown + End ( Poster session ) > link | 🔗 |
-
|
Bayesian Data Selection
(
Poster
)
>
|
Eli N Weinstein · Jeffrey Miller 🔗 |
-
|
Uncertainty estimation under model misspecification in neural network regression
(
Poster
)
>
|
Maria Cervera · Rafael Dätwyler · Francesco D'Angelo · Hamza Keurti · Benjamin F. Grewe · Christian Henning 🔗 |
-
|
Fast approximate BayesBag model selection via Taylor expansions
(
Poster
)
>
|
Neil Spencer · Jeffrey Miller 🔗 |
-
|
Diversity and Generalization in Neural Network Ensembles
(
Poster
)
>
|
Luis Antonio Ortega Andrés · Andres Masegosa · Rafael Cabañas 🔗 |
-
|
A shared parameter model accounting for drop-out not at random in a predictive model for systolic bloodpressure using the HUNT study
(
Poster
)
>
|
Aurora Christine Hofman 🔗 |
-
|
Influential Observations in Bayesian Regression Tree Models
(
Poster
)
>
|
Matthew Pratola 🔗 |
-
|
Invariant Priors for Bayesian Quadrature
(
Poster
)
>
|
Masha Naslidnyk · Javier González · Maren Mahsereci 🔗 |
-
|
Composite Goodness-of-fit Tests with Kernels
(
Poster
)
>
|
Oscar Key · Tamara Fernandez · Arthur Gretton · Francois-Xavier Briol 🔗 |
-
|
Inferior Clusterings in Misspecified Gaussian Mixture Models
(
Poster
)
>
|
Siva Rajesh Kasa · Vaibhav Rajan 🔗 |
-
|
Blindness of score-based methods to isolated components and mixing proportions
(
Poster
)
>
|
Li Kevin Wenliang · Heishiro Kanagawa 🔗 |
-
|
Bounding Wasserstein distance with couplings
(
Poster
)
>
|
Niloy Biswas · Lester Mackey 🔗 |
-
|
Relaxing the I.I.D. Assumption: Adaptively Minimax Optimal Regret via Root-Entropic Regularization
(
Poster
)
>
|
Blair Bilodeau · Jeffrey Negrea · Dan Roy 🔗 |
-
|
Measuring the sensitivity of Gaussian processes to kernel choice
(
Poster
)
>
|
Will Stephenson · Soumya Ghosh · Tin Nguyen · Mikhail Yurochkin · Sameer Deshpande · Tamara Broderick 🔗 |
-
|
Robust Bayesian Inference for Simulator-based Models via the MMD Posterior Bootstrap
(
Poster
)
>
|
Harita Dellaporta · Jeremias Knoblauch · Theodoros Damoulas · Francois-Xavier Briol 🔗 |
-
|
Your Bandit Model is Not Perfect: Introducing Robustness to Restless Bandits Enabled by Deep Reinforcement Learning
(
Poster
)
>
|
Jackson Killian · Lily Xu · Arpita Biswas · Milind Tambe 🔗 |
-
|
Bayesian Calibration of imperfect computer models using Physics-informed priors
(
Poster
)
>
|
Michail Spitieris 🔗 |
-
|
Forcing a model to be correct for classification
(
Poster
)
>
|
Jiae Kim · Steven MacEachern 🔗 |
-
|
Make cross-validation Bayes again
(
Poster
)
>
|
Yuling Yao · Aki Vehtari 🔗 |
-
|
Evaluating Bayesian Hierarchical Models for sc-RNA seq Data
(
Poster
)
>
|
Sijia Li 🔗 |
-
|
On Robustness of Counterfactuals in Structural Models
(
Poster
)
>
|
Yaroslav Mukhin 🔗 |
-
|
Robust Generalised Bayesian Inference for Intractable Likelihoods
(
Poster
)
>
|
Takuo Matsubara · Jeremias Knoblauch · Francois-Xavier Briol · Chris Oates 🔗 |
-
|
Boosting heterogeneous VAEs via multi-objective optimization
(
Poster
)
>
|
Adrián Javaloy · Maryam Meghdadi · Isabel Valera 🔗 |
-
|
PAC^m-Bayes: Narrowing the Empirical Risk Gap in the Misspecified Bayesian Regime
(
Poster
)
>
|
Joshua Dillon · Warren Morningstar · Alexander Alemi 🔗 |