Regularized Robustly Reliable Learners and Instance Targeted Attacks

TL;DR

This paper enhances robustly reliable learners for data poisoning attacks by introducing regularization to handle flexible hypothesis classes and developing faster algorithms using dynamic techniques.

Contribution

It proposes a regularized notion of robustly-reliable learners and introduces efficient algorithms with sublinear training time for certain cases.

Findings

Regularized learners provide meaningful guarantees for flexible hypothesis classes.

New algorithms achieve sublinear training time using dynamic techniques.

The approach improves robustness and efficiency in data poisoning scenarios.

Abstract

Instance-targeted data poisoning attacks, where an adversary corrupts a training set to induce errors on specific test points, have raised significant concerns. Balcan et al (2022) proposed an approach to addressing this challenge by defining a notion of robustly-reliable learners that provide per-instance guarantees of correctness under well-defined assumptions, even in the presence of data poisoning attacks. They then give a generic optimal (but computationally inefficient) robustly reliable learner as well as a computationally efficient algorithm for the case of linear separators over log-concave distributions. In this work, we address two challenges left open by Balcan et al (2022). The first is that the definition of robustly-reliable learners in Balcan et al (2022) becomes vacuous for highly-flexible hypothesis classes: if there are two classifiers h_0, h_1 \in H both with zero error on the training set such that h_0(x) \neq h_1(x), then a robustly-reliable learner must abstain on x. We address this problem by defining a modified notion of regularized robustly-reliable learners that allows for nontrivial statements in this case. The second is that the generic algorithm of Balcan et al (2022) requires re-running an ERM oracle (essentially, retraining the classifier) on each test point x, which is generally impractical even if ERM can be implemented efficiently. To tackle this problem, we show that at least in certain interesting cases we can design algorithms that can produce their outputs in time sublinear in training time, by using techniques from dynamic algorithm design.