CoChat

Feature Purification: How Adversarial Training Performs Robust Deep Learning

Open access
Allen-Zhu, Zeyuan, Li, Yuanzhi · 2022 · 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS)
Cited by 86

The DOI matches its Crossref record; 24 of the 24 references it cites were found in a registry or an index.

Abstract

Despite the empirical success of using adversarial training to defend deep learning models against adversarial perturbations, so far, it still remains rather unclear what the principles are behind the existence of adversarial perturbations, and what adversarial training does to the neural network to remove them. In this paper, we present a principle that we call feature purification, where we show one of the causes of the existence of adversarial examples is the accumulation of certain small dense mixtures in the hidden weights during the training process of a neural network; and more importantly, one of the goals of adversarial training is to remove such mixtures to purify hidden weights. We present both experiments on the CIFAR-10 dataset to illustrate this principle, and a theoretical result proving that for certain natural classification tasks, training a two-layer neural network with ReLU activation using randomly initialized gradient descent indeed satisfies this principle. Technically, we give, to the best of our knowledge, the first re-sult proving that the following two can hold simultaneously for training a neural network with ReLU activation. (1) Training over the original data is indeed non-robust to small adversarial perturbations of some radius. (2) Adversarial training, even with an empirical perturbation algorithm such as FGM, can in fact be provably robust against any perturbations of the same radius. Finally, we also prove a complexity lower bound, showing that low complexity models such as linear classifiers, low-degree polynomials, or even the neural tangent kernel for this network, cannot defend against perturbations of this same radius, no matter what algorithms are used to train them.11The full version of this paper can be found at https://arxiv.org/abs/-2005.l0190.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/focs52979.2021.00098

Cites

24 of 24 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 24 references

Cited by

See all 86

Metadata

Authors
Allen-Zhu, Z., & Li, Y.
Venue
2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS)
Publisher
IEEE
Published
2022-02
Pages
977-988
ISSN
0272-5428 (ISSN-L)
OpenAlex
W3028525609
Cited by
86
References
24
FWCI
8.09
Type
conference paper
Language
en
Open access
green

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Feature Purification: How Adversarial Training Performs Robust Deep Learning. https://app.cochat.ai/papers/10.1109/focs52979.2021.00098
The paper
Allen-Zhu, Z., & Li, Y. (2022). Feature Purification: How Adversarial Training Performs Robust Deep Learning. 2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS). https://doi.org/10.1109/focs52979.2021.00098

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.