Conference on

Mathematical and Statistical Challenges in Uncertainty Quantification

Scientific Organisers

Ismael Castillo (Paris, Sorbonne Université)

Richard Nickl (University of Cambridge)

Judith Rousseau (University of Oxford)

COVID 19 UPDATE: The originally planned conference had to be cancelled. A smaller scale virtual conference was hosted via zoom.

Scientific Goals: Topics will broadly include foundational questions and mathematical challenges involving (Bayesian and non-Bayesian) approaches to the construction of inference procedures in high-dimensional and complex nonparametric statistical models arising in modern data science.

This conference is funded by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 647812, UQMS I).

Virtual Zoom Conference, July 2020

Here is a video-link to recordings of the talks

Programme

July 13 (all times BST; chair: A.W. van der Vaart)

9:00-10:00 Gabriel Paternain (Cambridge); Statistical aspects of non-Abelian X-ray transforms

10:20-11:00 Kweku Abraham (Paris-Saclay); Controlling the false discovery rate of an empirical Bayes multiple testing procedure

11:00-11:40 Amine Hadji (Leiden); Distributed methods for Bayesian regression: contraction rate and uncertainty quantification

July 14 (all times BST; chair: I. Castillo)

9:00-10:00 Botond Szabo (Leiden); On distributed estimation and testing under communication constraints

10:20-11:00 Sven Wang (Cambridge); On polynomial-time computation of high-dimensional posterior measures by Langevin-type algorithms

11:00-11:40 Deborah Sulem (Oxford); Bayesian estimation of nonlinear Hawkes processes

July 15 (all times BST; chair: J. Rousseau)

9:00-10:00 Etienne Roquain (Paris); Sparse multiple testing: can one estimate the null distribution?

10:20-11:00 Kolyan Ray (Imperial); Variational Bayes for high-dimensional linear regression with sparse priors

11:00-11:40 Thibault Randrianarisoa (Paris); A toy model of Polya tree ensemble: smoothing and adaptation

July 16 (all times BST; chair: R. Nickl)

9:00-10:00 Judith Rousseau (Oxford); Posterior concentration and Bernstein-von Mises theorems for deconvolution models.

10:20-11:00 Randolf Altmeyer (Cambridge); Towards a Bernstein-von Mises theorem for SPDEs

11:00-11:40 Stefan Franssen (Leiden); Empirical Bayesian uncertainty quantification using deep neural network regression in Besov spaces