High-Dimensional Structure Theory Team Seminar(Talk by Christophe Quentin Valvason, University of Geneva).

The High-Dimensional Structure Theory Team at RIKEN AIP hosts a one-hour online talk by Christophe Quentin Valvason of the University of Geneva, titled "Optimal Transport for Nonparametric Inference: Improving Confidence Intervals". The talk presents a method that recasts confidence interval construction as an optimal transport problem between a discrete source measure and an estimate of the sampling distribution. It covers finite-sample bounds on non-coverage probability, asymptotic validity including consistency in the Hausdorff metric, and a regularization view of why finite support stabilizes the intervals. Simulations show coverage closer to the nominal level than standard bootstrap procedures, which matters most in small samples.