[Statlist] Research Seminar in Statistics *FRIDAY, 6 MAY 2022* GSEM, University of Geneva

[gsem-support-instituts]

Dear All,

We are pleased to invite you to our next Research Seminar.

Looking forward to seeing you,


Organized by Professor Sebastian Engelke on behalf of the Research Center for Statistics (https://www.unige.ch/gsem/en/research/institutes/rcs/)


FRIDAY, 6 MAY 2022 at 11:15am, Uni-Mail M 5220 & ONLINE
Zoom research webinar: https://unige.zoom.us/j/92924332087?pwd=U1U1NFk4dTFCRHBMeWYrSDBQcXBiQT09
Meeting ID: 929 2433 2087
Passcode: 399192

Limiting Laws for Optimal Transport Plans on Finite Spaces
Yoav ZEMEL, University of Cambridge, UK
https://www.dpmms.cam.ac.uk/person/yz668

ABSTRACT:
Optimal transport is now a popular tool in statistics, machine learning, and data science. The majority of studies regarding the asymptotic properties of optimal transport have focused on the Wasserstein distance itself, i.e., the optimal objective value. In many situations, however, it is the transport plan (or map) that is more informative, as it allows the practitioner to understand \emph{where}, and not only how much, transport is taking place. We thus study the asymptotics of optimal transport plans in the (practically relevant) case of finite ground space. Possible limiting distributions are derived, and it is shown that the limiting distributions are non-standard with complexity that depends on the degeneracy of the optimal transport linear program and its dual. In particular, if the dual is degenerate (as it is in regular ground spaces), the asymptotic distribution depends on the way the empirical optimal solution is chosen. The results are not specific to optimal transport and hold for general linear programs.


Visit the website: https://www.unige.ch/gsem/en/research/seminars/rcs/