NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:9114
Title:Inherent Tradeoffs in Learning Fair Representations


		
This paper shows information-theoretic lower bounds characterizing fairness-utility trade-offs in representation learning. The work is interesting, novel and timely, and has the potential to inspire new research directions in fairness in ML.