BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Supervised Learning without Discrimination
DTSTART:20170616T160000
DTEND:20170616T164500
DTSTAMP:20260408T101704Z
UID:a9d2a00b40cf2fc06ce37de7dd9a18a32c79375fd95cdafeda0b6a03
CATEGORIES:Conferences - Seminars
DESCRIPTION:Nathan Srebro\, Toyota Technological Institute at Chicago\nAs 
 machine learning is increasingly being used in areas protected by anti dis
 crimination law\, or in other domains which are socially and morally sensi
 tive\, the problem of algorithmically measuring and avoiding prohibited di
 scrimination in machine learning is pressing. What does it mean for a pred
 ictor to not discriminate with respect to protected group (e.g. according 
 to race\, gender\, etc)? We propose a notion of non-discrimination that ca
 n be measured statistically\, used algorithmically\, and avoids many of th
 e pitfalls of previous definitions. We further study what type of discrimi
 nation and non-discrimination can be identified with oblivious tests\, whi
 ch treat the predictor as an opaque black-box\, and what different oblivio
 us tests tell us about possible discrimination. Joint work with Suriya Gun
 asekar\, Mortiz Hardt\, Mesrob Ohannessian\, Eric Pierce and Blake Woodwoo
 rth.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR