On fairness, discrimination and conditional parity as a notion of non-discrimination in machine learning

4 luglio 2018
Versione stampabile

Time: 11.00 to 12.00 
Location: Garda room, Via Sommarive 5 - Polo Ferrari 1 (Povo, TN) 

Speaker

Prof. Ya’acov Ritov, University of Michigan

Abstract

We identify conditional parity as a general notion of nondiscrimination in machine learning. In fact, several recently proposed notions of non-discrimination, including a few counterfactual notions, are instances of conditional parity. We show that conditional parity is amenable to statistical analysis by studying randomization as a general mechanism for achieving conditional parity and a kernel-based test of conditional parity.

About the Speaker

Ya'acov Ritov has B.Sc. and M.Sc in Electrical Engineering from the Technion and PhD in statistics from the Hebrew University, where he was in the department of statistics until 2015. He is currently a professor of statistics at the University of Michigan

Contact Person Regarding this Talk: andrea.passerini [at] unitn.it (Andrea Passerini )