BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Understanding Generalization in Adaptive Data Analysis
DTSTART:20170616T140000
DTEND:20170616T144500
DTSTAMP:20260407T055627Z
UID:1ce058a599f8e2af1a6bbbe446c952b855773743329077fccec284b3
CATEGORIES:Conferences - Seminars
DESCRIPTION:Vitaly Feldman\, IBM Research\nDatasets are often reused to pe
 rform multiple statistical analyses in an adaptive way\, in which each ana
 lysis may depend on the outcomes of previous analyses on the same dataset.
  Standard statistical guarantees do not account for these dependencies and
  little is known about how to provably avoid overfitting in the adaptive s
 etting. In this talk I'll describe a new framework to address this problem
 \, an approach based on differential privacy\, and several algorithms base
 d on this approach. Based on joint works with Dwork\, Hardt\, Pitassi\, Re
 ingold and Roth (STOC 2015\, NIPS 2015) and with Steinke (COLT 2017)
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR