BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20251008T095215EDT-2910sSmfaE@132.216.98.100 DTSTAMP:20251008T135215Z DESCRIPTION:Title: The HulC: Hull based Confidence Regions.\n\n\n Abstract: \n\n\nWe develop and analyze the HulC\, an intuitive and general method fo r constructing confidence sets using the convex hull of estimates construc ted from subsets of the data. Unlike classical methods which are based on estimating the (limiting) distribution of an estimator\, the HulC is often simpler to use and effectively bypasses this step. In comparison to the b ootstrap\, the HulC requires fewer regularity conditions and succeeds in m any examples where the bootstrap provably fails. Unlike subsampling\, the HulC does not require knowledge of the rate of convergence of the estimato rs on which it is based. The validity of the HulC requires knowledge of th e (asymptotic) median-bias of the estimators. We further analyze a variant of our basic method\, called the Adaptive HulC\, which is fully data-driv en and estimates the median-bias using subsampling. We show that the Adapt ive HulC retains the aforementioned strengths of the HulC. In certain case s where the underlying estimators are pathologically asymmetric\, the HulC and Adaptive HulC can fail to provide useful confidence sets. We discuss these methods in the context of several challenging inferential problems w hich arise in parametric\, semi-parametric\, and non-parametric inference. Although our focus is on validity under weak regularity conditions\, we a lso provide some general results on the width of the HulC confidence sets\ , showing that in many cases the HulC confidence sets have near-optimal wi dth. Please let me know if you need anything else.\n\n\n Speaker\n\n\nArun Kumar is an Assistant Professor at the Department of Statistics and Data S cience\, Carnegie Mellon University. He graduated from the Wharton School of the University of Pennsylvania on May 17\, 2020 with a Ph.D. in Statist ics. His advisors are Lawrence D. Brown and Andreas Buja.\n\nHis research interests include post-selection inference\, large sample theory\, robust statistics\, semi-parametric statistics\, non-parametric statistics\, conc entration inequalities\, high-dimensional CLT\, and dependent data.\n\n\n \n \n https://mcgill.zoom.us/j/83436686293?pwd=b0RmWmlXRXE3OWR6NlNIcWF5d0dJQ T09\n\n Meeting ID: 834 3668 6293\n\n Passcode: 12345\n\n  \n\n  \n \n \n\n DTSTART:20211001T193000Z DTEND:20211001T203000Z SUMMARY:Arun Kumar (Carnegie Mellon University) URL:/mathstat/channels/event/arun-kumar-carnegie-mello n-university-333743 END:VEVENT END:VCALENDAR