[Statlist] Reminder: ETH-FDS Seminar talk with Sven Wang, EPFL, Lausanne -- 5 March 2026, 16:15, ETH Zurich

[Maurer Letizia]

We are pleased to announce and invite you to the following talk in our ETH-FDS seminar series:

„On global polynomial-time computable estimators in statistical nonlinear inverse problems“

by Sven Wang, EPFL, Lausanne

Date and Time (Zurich):  Thursday, 5 March 2026, 16:15 - 17.15
Place: 				HG E 5

Abstract: Non-linear statistical inverse problems pose major challenges both for statistical analysis and computation. Likelihood-based estimators typically lead to non-convex and possibly multimodal optimization landscapes, and Markov chain Monte Carlo (MCMC) methods may mix exponentially slowly. We discuss recent progress in devising both statistical and global polynomial-time computational guarantees in such settings. In particular, we will discuss a class of computationally tractable estimators--plug-in and PDE-penalized M-estimators--for inverse problems defined through parametric PDEs. These estimators arise from conditionally convex and, in many PDE examples, nested quadratic optimization formulations which avoid evaluating the forward map G(f) and do not require PDE solvers. For prototypical non-linear inverse problems arising from elliptic PDEs, such as the well-known Darcy model, we prove that these estimators attain the best currently known statistical convergence rates while being globally computable in polynomial time. In the Darcy model, we obtain novel sub-quadratic o(N^2) arithmetic runtime bound for estimating f from N noisy samples. Our analysis is based on new generalized stability estimates, extending classical stability beyond the range of the forward operator, combined with tools from nonparametric M-estimation. Our estimators also provide principled warm-start initializations for polynomial-time Bayesian computation. Based on the preprint https://arxiv.org/abs/2601.09007

Seminar websites: https://math.ethz.ch/sfs/news-and-events/data-science-seminar.html, https://math.ethz.ch/sfs/eth-foundations-of-data-science.html

Organisers: A. Bandeira, H. Bölcskei, P. Bühlmann, J. Peters, F. Yang



ETH Zürich I Seminar für Statistik I Rämistrasse 101 I 8092 Zürich I Telefon +41 44 632 65 18 I sekretariat at stat.math.ethz.ch