Seminars and Events

Artificial Intelligence Seminar

Fair Learning with Private Demographic Data

Event Details

Sensitive attributes such as race are rarely available to learners in real world settings, as their collection is often restricted by laws and regulations. We give a scheme that allows individuals to release their sensitive information privately while still allowing any downstream entity to learn non-discriminatory predictors. We show how to adapt non-discriminatory learners to work with privatized protected attributes giving theoretical guarantees on performance. Finally, we highlight how the methodology could apply to learning fair predictors in settings where protected attributes are only available for a subset of the data. (Joint work with Hussein Mozannar, MIT, and Nathan Srebro, TTIC.)

TALK NOT RECORDED 

Speaker Bio

Mesrob I. Ohannessian is an Assistant Professor at the ECE department of the University of Illinois at Chicago. He was previously a Research Assistant Professor at the Toyota Technological Institute at Chicago. He held postdoctoral positions at UCSD, MSR-Inria, and Université Paris-Sud. He received his PhD in EECS from MIT. His research interests are in machine learning, statistics, information theory, and their applications, particularly to problems marked by data scarcity and to decisions that affect society.