BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Preventing Fairness Gerrymandering: Auditing and Learning for Subg
 roup Fairness - Prof. Michael Kearns (University of Pennsylvania)
DTSTART:20180614T100000Z
DTEND:20180614T110000Z
UID:TALK106339@talks.cam.ac.uk
CONTACT:Adrian Weller
DESCRIPTION:The most prevalent notions of fairness in machine learning are
  statistical and coarse: they fix a small collection of pre-defined groups
  or attributes (such as race or gender)\, and then ask for parity of some 
 statistic of the classifier (such as false positive rate) across these gro
 ups. Constraints of this form are susceptible to intentional or inadverten
 t "fairness gerrymandering"\, in which a classifier appears to be fair on 
 each individual group\, but badly violates fairness on one or more subgrou
 ps defined over the protected attributes. \n\nWe propose to instead demand
  notions of fairness across exponentially (or infinitely) many subgroups\,
  defined by a structured class of functions over the protected attributes.
   This interpolates between statistical definitions of fairness and recent
 ly proposed individual notions of fairness. While the problem of auditing 
 a given classifier for subgroup fairness can be computationally intractabl
 e in the worst case\, we show that it can itself be cast as an instance of
  weighted classification\, and thus standard learning algorithms can be ap
 plied.\n\nWe then present algorithms that provably converge to the best su
 bgroup-fair classifier. The algorithms are based on a formulation of subgr
 oup fairness as a two-player zero-sum game between a Learner and an Audito
 r. We present extensive empirical validation on a number of datasets in wh
 ich fairness is a concern\, and demonstrate that appealing trade-offs betw
 een accuracy and subgroup fairness are possible in practice.\n\nJoint work
  with Seth Neel\, Aaron Roth\, and Zhiwei Steven Wu.
LOCATION:Cambridge University Engineering Department\, Lecture Room 4
END:VEVENT
END:VCALENDAR