Multigroup Robustness

TL;DR

This paper introduces multigroup robust algorithms that provide tailored robustness guarantees for different subpopulations, addressing localized data corruption and linking fairness with robustness in machine learning models.

Contribution

It proposes a novel approach to robustness that considers subpopulation-specific data corruption, bridging multigroup fairness and robustness guarantees.

Findings

Algorithms offer robustness guarantees that degrade with subpopulation-specific corruption.

The approach outperforms standard methods when corruption is unevenly distributed.

Establishes a new connection between multigroup fairness and robustness.

Abstract

To address the shortcomings of real-world datasets, robust learning algorithms have been designed to overcome arbitrary and indiscriminate data corruption. However, practical processes of gathering data may lead to patterns of data corruption that are localized to specific partitions of the training dataset. Motivated by critical applications where the learned model is deployed to make predictions about people from a rich collection of overlapping subpopulations, we initiate the study of multigroup robust algorithms whose robustness guarantees for each subpopulation only degrade with the amount of data corruption inside that subpopulation. When the data corruption is not distributed uniformly over subpopulations, our algorithms provide more meaningful robustness guarantees than standard guarantees that are oblivious to how the data corruption and the affected subpopulations are related. Our techniques establish a new connection between multigroup fairness and robustness.