Certified Robustness to Data Poisoning in Gradient-Based Training

TL;DR

This paper introduces a novel framework that provides provable guarantees on the robustness of gradient-trained models against data poisoning and backdoor attacks, without altering the training process.

Contribution

It develops the first method to certify model robustness against various poisoning threats using convex relaxations and parameter set over-approximations.

Findings

Certifies robustness for untargeted and targeted poisoning attacks.

Provides bounds on model performance and backdoor success rate.

Demonstrates effectiveness on real-world datasets.

Abstract

Modern machine learning pipelines leverage large amounts of public data, making it infeasible to guarantee data quality and leaving models open to poisoning and backdoor attacks. Provably bounding model behavior under such attacks remains an open problem. In this work, we address this challenge by developing the first framework providing provable guarantees on the behavior of models trained with potentially manipulated data without modifying the model or learning algorithm. In particular, our framework certifies robustness against untargeted and targeted poisoning, as well as backdoor attacks, for bounded and unbounded manipulations of the training inputs and labels. Our method leverages convex relaxations to over-approximate the set of all possible parameter updates for a given poisoning threat model, allowing us to bound the set of all reachable parameters for any gradient-based learning algorithm. Given this set of parameters, we provide bounds on worst-case behavior, including model performance and backdoor success rate. We demonstrate our approach on multiple real-world datasets from applications including energy consumption, medical imaging, and autonomous driving.