The work Professor Wang presents at this seminar will combine ideas from independence testing, spatial process and optimal transport, and it will be accessible to a broader audience.
Title
Coverage correlation: detecting singular dependencies between random variables
Abstract
We introduce the coverage correlation coefficient, a novel, symmetric, nonparametric measure of statistical association designed to detect dependencies concentrated on low-dimensional structures within the joint distribution of two random variables or vectors. Based on Monge-Kantorovich ranks and geometric coverage processes, our statistic quantifies the extent to which the joint distribution deviates from independence by measuring the uncovered volume in a high-dimensional unit cube. The coverage correlation coefficient is distribution-free, admits an analytically tractable asymptotic null distribution, and can be computed efficiently, making it well-suited for uncovering complex, potentially nonlinear association large-scale pairwise testing.
About
Tengyao Wang, professor at the Department of Statistics at London School of Economics, is visiting UiO. At this joint seminar between Integreat and Department of Mathematics he will talk about his most recent results. Tengyao Wang is broadly interested in high dimensional statistics and highly productive. He is awarded both the RSS Research Prize and Guy Medal, and is known for his inspiring talks.
Practical
For those unable to attend in person: https://uio.zoom.us/j/62535744013
Point of contacts:
- Organiser and moderator: Professor Ingrid K. Glad, University of Oslo
- Integreat Administrative Leader Maria Dikova, University of Oslo