BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Jose Blanchet (Stanford University)
DTSTART:20201023T150500Z
DTEND:20201023T160500Z
DTSTAMP:20260423T005728Z
UID:sss/10
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/sss/10/">Sta
 tistical Aspects of Wasserstein Distributionally Robust Optimization Estim
 ators</a>\nby Jose Blanchet (Stanford University) as part of Stochastics a
 nd Statistics Seminar Series\n\n\nAbstract\nAbstract: Wasserstein-based di
 stributional robust optimization problems are formulated as min-max games 
 in which a statistician chooses a parameter to minimize an expected loss a
 gainst an adversary (say nature) which wishes to maximize the loss by choo
 sing an appropriate probability model within a certain non-parametric clas
 s. Recently\, these formulations have been studied in the context in which
  the non-parametric class chosen by nature is defined as a Wasserstein-dis
 tance neighborhood around the empirical measure. It turns out that by appr
 opriately choosing the loss and the geometry of the Wasserstein distance o
 ne can recover a wide range of classical statistical estimators (including
  Lasso\, Graphical Lasso\, SVM\, group Lasso\, among many others). This ta
 lk studies a wide range of rich statistical quantities associated with the
 se problems\; for example\, the optimal (in a certain sense) choice of the
  adversarial perturbation\, weak convergence of natural confidence regions
  associated with these formulations\, and asymptotic normality of the DRO 
 estimators. (This talk is based on joint work with Y. Kang\, K. Murthy\, a
 nd N. Si.)\n
LOCATION:https://researchseminars.org/talk/sss/10/
END:VEVENT
END:VCALENDAR