Request For Research Presentations For the PrivacyCon Conference #00032

Submission Number:
00032
Commenter:
Daniel Hsu
State:
New York
Initiative Name:
Request For Research Presentations For the PrivacyCon Conference
This submission is a research article by Florian Tramèr (EPFL), Vaggelis Atlidakis (Columbia), Roxana Geambasu (Columbia), Daniel Hsu (Columbia), Jean-Pierre Hubaux (EPFL), Mathias Humbert (EPFL), Ari Juels (Cornell Tech), and Huang Lin (EPFL). We describe FairTest, a testing toolkit that detects unwarranted associations between a data-driven algorithm's outputs (e.g., prices or labels) and user subpopulations, including sensitive groups (e.g., defined by race or gender). FairTest reports statistically significant associations to programmers as association bugs, ranked by their strength and likelihood of being unintentional, rather than necessary effects.