Testing Most Influential Sets

Published: 30 Oct 2025, Last Modified: 30 Oct 2025NetSciX 2026 PosterEveryoneRevisionsBibTeXCC BY-SA 4.0

Keywords: influence, data attribution, model sensitivity, fairness, interpretable

TL;DR: We characterize the distribution of influence from most influential sets, providing theoretical foundations for hypothesis tests

Submission Number: 169

Loading