[AIP Seminar] Martin Wainwright: Wild refitting for black box prediction

This talk will be held in a hybrid format, both in person at Open Space* of RIKEN AIP (Nihonbashi office) and online by Zoom. Open Space: *only available to AIP researchers.

DATE, TIME & LOCATION
Friday, August 29th, 11:00 - 12:30, RIKEN AIP Nihombashi Office, Open Space

TITLE
Wild refitting for black box prediction

ABSTRACT
We describe a novel procedure for estimating the excess risk of a prediction method for regression. Inspired by the wild bootstrap, it uses Rademacher symmetrization to construct a synthetic dataset for refitting. Unlike the bootstrap, it requires only a single refit, and we give non-asymptotic guarantees on the risk estimate. Notably, the method requires only black-box access to the predictor, meaning that no understanding of its internals is required; and it does not require any hold-out or cross-validation, so that it can be applied to non-i.i.d. or heterogeneous datasets. We illustrate its behavior for non-rigid structure-from-motion, and plug-and-play image denoising using deep net priors.

Pre-print: https://arxiv.org/abs/2506.21460

BIO
Martin Wainwright is the Ford Professor in Electrical Engineering and Computer Science and Mathematics at MIT, and affiliated with the Laboratory for Information and Decision Systems and Statistics and Data Science Center. He is broadly interested in statistics, machine learning, information theory and algorithms. He has received a number of awards and recognition including a John Simon Guggenheim Fellowship, Alfred P. Sloan Foundation Fellowship, the COPSS Presidents’ Award from the Joint Statistical Societies, a Section Lecturer with the International Congress of Mathematicians, and the Blackwell Lectureship and Award from the Institute of Mathematical Statistics.