Saved in:
Bibliographic Details
Main Author: Bahmani, Sohail
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2411.05853
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • We derive a fundamental trade-off between standard and adversarial risk in a rather general situation that formalizes the following simple intuition: "If no (nearly) optimal predictor is smooth, adversarial robustness comes at the cost of accuracy." As a concrete example, we evaluate the derived trade-off in regression with polynomial ridge functions under mild regularity conditions. Generalizing our analysis of this example, we formulate a necessary condition under which adversarial robustness can be achieved without significant degradation of the accuracy. This necessary condition is expressed in terms of a quantity that resembles the Poincaré constant of the data distribution.