Abstract
dc:description.abstractImplicit bias is the unconscious attribution of particular qualities (or lack of) to a member from a particular social group (e.g. defined by race or gender). Studies on implicit bias have shown that these unconscious stereotypes can have adverse outcomes in various social contexts, such as job screening, teaching, or policing. This dissertation advocates for an application of fairness based re-ranking methods to improve the fairness to all items which, to some surprise, comes with little cost to or can even improve the utility. We present our key contributions in ranking when in the presence of implicit bias. This includes the development of a theorem where we prove that under simplifying assumptions on the utilities of items, simple, well-studied, constraints can ensure that the utility does not decrease with respect to a naive ranking. Finally, we augment our theoretical results with empirical findings on real-world distributions from the IIT-JEE (2009) dataset.
Degree
thesis:*- Name thesis:degree_name
- M.S., Computer Science
- Level thesis:degree_level
- Master of Science
- Discipline thesis:degree_discipline
- Computer Science
- Grantor
- Cornell University
- Year dc:date.issued
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Bayoumi, Magd
- Chair dc:contributor.chair
-
- Joachims, Thorsten
- Committee member dc:contributor.committeemember
-
- Cardie, Claire
Subjects
dc:subject × 3Rights
dc:rights- Statement dc:rights
-
- Attribution-NoDerivatives 4.0 International
- Licence dc:rights.uri
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Bibid
- 15644158
- Dc Identifier Other
-
ProQuest Submission ID: 11599
ProQuest Publication ID: 29996238