BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//ox.ac.uk//NONSGML oxford.event//EN
X-WR-TIMEZONE:Europe/London
BEGIN:VTIMEZONE
TZID:Europe/London
X-LIC-LOCATION:Europe/London
BEGIN:DAYLIGHT
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:BST
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:GMT
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Distinguished Speaker Seminar
DTSTART;TZID=Europe/London:20260528T153000
DTEND;TZID=Europe/London:20260528T163000
DTSTAMP:20260527T003719Z
UID:3e565fcb-d455-f111-bec7-7ced8d9a5614
CREATED:20260522T115312Z
DESCRIPTION:Title:        Distribution-Free Nonparametric Inference Based 
 on Optimal Transport and Kernel Methods \nAbstract: The Wilcoxon rank-sum 
 (or Mann–Whitney) test is one of the most widely used tools for comparin
 g two groups without making assumptions about the underlying data distribu
 tion. One of the reasons for its enduring popularity is a remarkable resul
 t of Hodges and Lehmann (1956)\, which shows that the asymptotic relative 
 efficiency of Wilcoxon's test with respect to Student's t-test\, under loc
 ation alternatives\, never falls below 0.864\, despite the former being di
 stribution-free in finite samples. Even more striking is the result of Che
 rnoff and Savage (1958)\, which shows that the efficiency of a Gaussian sc
 ore transformed Wilcoxon's test\, against the t-test\, is lower bounded by
  1. In other words\, the Gaussian score transformed Wilcoxon test uniforml
 y dominates the t-test in terms of efficiency\, while also remaining distr
 ibution-free.\n\nIn this talk we will discuss multivariate versions of the
 se celebrated results\, by considering distribution-free analogues of the 
 Hotelling T²-test based on optimal transport. The proposed tests are cons
 istent against a general class of alternatives and satisfy Hodges-Lehmann 
 and Chernoff-Savage-type efficiency lower bounds over various natural fami
 lies of multivariate distributions\, despite being entirely agnostic to th
 e underlying data generating mechanism. We will also discuss how optimal-t
 ransport-based multivariate ranks can be used to construct distribution-fr
 ee analogues of the celebrated kernel two-sample test that enjoy a trifect
 a of desirable properties: universal consistency\, efficient computation\,
  and nontrivial asymptotic efficiency. \n\nShort Bio: Bhaswar B. Bhattacha
 rya is an Associate Professor in the Department of Statistics and Data Sci
 ence at the Wharton School\, University of Pennsylvania. He received his P
 h.D. from the Department of Statistics at Stanford University in 2016. Pri
 or to that\, he obtained his Bachelor and Master degrees in Statistics fro
 m the Indian Statistical Institute\, Kolkata in 2009 and 2011\, respective
 ly. His research interests include non-parametric statistics\, combinatori
 al probability\, and discrete and computational
LAST-MODIFIED:20260522T121206Z
LOCATION:Department of Statistics - Large Lecture Theatre\, Large Lecture 
 Theatre Department of Statistics 24-29 St Giles' Oxford Oxfordshire OX1 3L
 B United Kingdom
SPEAKER:Professor Bhaswar B. Bhattacharya (Department of Statistics and Da
 ta Science\, Wharton School\, University of Pennsylvania)
END:VEVENT
END:VCALENDAR