Welcome!

Bayesian inference has become a popular framework for decision-making given its consistent and flexible handling of uncertainty. In this regime, however, the statistician is subject to several surprisingly strong assumptions, which are violated in almost all modern machine learning settings. This is in fact well-understood, and has led to a range of methods which aim to retain characteristics of Bayesian uncertainty quantification without the restrictive assumptions that underpin it. Collectively, this body of work is sometimes referred to as “generalised Bayes”. This name, however, does not capture the main appeal of these conceptual frameworks: by unapologetically endorsing posteriors that lie outside the confines of Bayesian epistemology, they are intrinsically post-Bayesian. This is not a minor difference in semantics, but a major shift in outlook.

The first workshop on Advances in post-Bayesian methods aims to bring together the currently disparate subfields, stretching from PAC Bayes, generalised Bayes, predictive resampling, and Martingale posteriors, to online learning and beyond. With each subfield growing faster now than ever before, it is imperative that we bring them together to form a unified post-Bayesian front. And given the growing public interest in probabilistic AI, aligning the direction of our field is of especially critical importance. Over the course of two days, we will host eight invited talks from leaders across the post-Bayesian landscape, eight contributed talks, and a poster session to ignite discussion and innovation in our growing community, and will compliment the post-Bayesian seminar series.

Location

All of the sessions will be held in the Denys Holland Lecture Theatre in Bentham House at UCL. The nearest tube stations are Euston, Warren Street, King’s Cross, and Russell Square.

Google Maps | what3words

Key dates

Programme

All presenters are listed in alphabetical order.

Talks

Keynote speakers

Dr. Badr-Eddine Chérief-Abdellatif (Sorbonne Université)
On the Asymptotics of Importance Weighted Variational Inference (abstract)

Dr. Diana Cai (Flatiron Institute)
Batch and match: score-based approaches for black-box variational inference (abstract)

Dr. Kamélia Daudel (ESSEC Business School)
Learning with Importance Weighted Variational Inference (abstract)

Prof. Peter Gruenwald (CWI)
E-Posteriors, Tempered Posteriors, Test Supermartingales (abstract)

Prof. Benjamin Guedj (UCL, Alan Turing Institute, INRIA)
Comparing Comparators in Generalization Bounds (abstract)

Dr. Takuo Matsubara (University of Edinburgh)
Wasserstein Gradient Boosting: A Framework for Distribution-Valued Supervised Learning (abstract)

Prof. Sonia Petrone (Bocconi University)
Bayesian predictive inference (abstract)

Dr. Susan Wei (Monash University)
What Is Prediction For? Turning Pre-Trained Transformers into Bayesian Engines (abstract)

Contributed talks

Dr. Sergio Bacallado (University of Cambridge)
Generalisation bounds for predictive risk (abstract)

Dr. Pier Giovanni Bissiri (University of Milano-Bicocca)
A new look at fiducial inference (abstract)

Prof. François Caron (University of Oxford)
Confidence with a Robust Bayesian Touch (abstract)

Ioar Casado-Telletxea (Basque Center for Applied Mathematics)
A PAC-Bayesian Perspective on Marginal Likelihood and Generalization (abstract)

Dr. Lorenzo Cappello (Universitat Pompeu Fabra)
New (and old) predictive schemes with “a.c.i.d.” sequences (abstract)

Prof. Imma Curato (TU Chemnitz)
Mixed moving average field guided learning for spatio-temporal data (abstract)

Dr. Harita Dellaporta (UCL)
Decision Making under Model Misspecification: DRO with Robust Bayesian Ambiguity Sets (abstract)

Valentin Kilian (University of Oxford)
Confidence sequences with informative, bounded-influence priors (abstract)

Kai Lehman (LMU)
Post-Bayesian Inference for Misspecified Cosmological Models (abstract)

Dr. Yingzhen Li (Imperial)
Variational uncertainty decomposition for LLM in-context learning (abstract)

Devina Mohan (University of Manchester)
Challenges in Practical Bayesian Deep Learning: A Case Study in Radio Galaxy Classification (abstract)

Dr. Sam Power (Bristol University)
Gradient Flows for Statistical Computation - Trends and Trajectories (abstract)

Prof. Martyn Plummer (Warwick University)
Bayesian estimating equations (abstract)

Ruchira Ray (Columbia University)
Statistical guarantees for data-driven posterior tempering (abstract)

Mahalakshmi Sabanayagam (Technical University of Munich)
Generalization Certificates for Adversarially Robust Bayesian Linear Regression (abstract)

Alisa Sheinkman (University of Edinburgh)
Variational bow tie neural networks with shrinkage (abstract)

Dr. Zheyang Shen (Newcastle University)
Prediction-Centric Uncertainty Quantification via MMD (abstract)

Dr. Yuli Slavutsky (Columbia University)
Quantifying Uncertainty in the Presence of Distribution Shifts (abstract)

Posters

Organisers

Dr. Jeremias Knoblauch (UCL)
Yann McLatchie (UCL)
Matías Altamirano (UCL)

Sponsors