Poster
On preserving non-discrimination when combining expert advice
Avrim Blum · Suriya Gunasekar · Thodoris Lykouris · Nati Srebro

Wed Dec 5th 10:45 AM -- 12:45 PM @ Room 517 AB #136
in Wed Poster Session A »

We study the interplay between sequential decision making and avoiding discrimination against protected groups, when examples arrive online and do not follow distributional assumptions. We consider the most basic extension of classical online learning: Given a class of predictors that are individually non-discriminatory with respect to a particular metric, how can we combine them to perform as well as the best predictor, while preserving non-discrimination? Surprisingly we show that this task is unachievable for the prevalent notion of "equalized odds" that requires equal false negative rates and equal false positive rates across groups. On the positive side, for another notion of non-discrimination, "equalized error rates", we show that running separate instances of the classical multiplicative weights algorithm for each group achieves this guarantee. Interestingly, even for this notion, we show that algorithms with stronger performance guarantees than multiplicative weights cannot preserve non-discrimination.

Keywords: Online Learning · Fairness, Accountability, and Transparency

Author Information

Avrim Blum (Toyota Technological Institute at Chicago)
Suriya Gunasekar (TTI Chicago)
Thodoris Lykouris (Cornell University)
Nati Srebro (TTI-Chicago)

More from the Same Authors