MIMIR: Masked Image Modeling for Mutual Information-based Adversarial Robustness Article Swipe

MAKE IMPACT

Xiaoyun Xu , Shujian Yu , Jingzheng Wu , Stjepan Picek · YOU? · · 2023 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2312.04960

Vision Transformers (ViTs) have emerged as a fundamental architecture and serve as the backbone of modern vision-language models. Despite their impressive performance, ViTs exhibit notable vulnerability to evasion attacks, necessitating the development of specialized Adversarial Training (AT) strategies tailored to their unique architecture. While a direct solution might involve applying existing AT methods to ViTs, our analysis reveals significant incompatibilities, particularly with state-of-the-art (SOTA) approaches such as Generalist (CVPR 2023) and DBAT (USENIX Security 2024). This paper presents a systematic investigation of adversarial robustness in ViTs and provides a novel theoretical Mutual Information (MI) analysis in its autoencoder-based self-supervised pre-training. Specifically, we show that MI between the adversarial example and its latent representation in ViT-based autoencoders should be constrained via derived MI bounds. Building on this insight, we propose a self-supervised AT method, MIMIR, that employs an MI penalty to facilitate adversarial pre-training by masked image modeling with autoencoders. Extensive experiments on CIFAR-10, Tiny-ImageNet, and ImageNet-1K show that MIMIR can consistently provide improved natural and robust accuracy, where MIMIR outperforms SOTA AT results on ImageNet-1K. Notably, MIMIR demonstrates superior robustness against unforeseen attacks and common corruption data and can also withstand adaptive attacks where the adversary possesses full knowledge of the defense mechanism.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2312.04960

PDF

https://arxiv.org/pdf/2312.04960

OA Status

green

Cited By

1

Related Works

10

OpenAlex ID

https://openalex.org/W4389599655

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4389599655

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2312.04960

Digital Object Identifier

Title

MIMIR: Masked Image Modeling for Mutual Information-based Adversarial Robustness

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2023

Year of publication

Publication date

2023-12-08

Full publication date if available

Authors

Xiaoyun Xu, Shujian Yu, Jingzheng Wu, Stjepan Picek

List of authors in order

Landing page

https://arxiv.org/abs/2312.04960

Publisher landing page

PDF URL

https://arxiv.org/pdf/2312.04960

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2312.04960

Direct OA link when available

Concepts

Adversarial system, Computer science, Artificial intelligence, Machine learning, Adversary, Bottleneck, Convolutional neural network, Mutual information, Robustness (evolution), Computer security, Biochemistry, Embedded system, Chemistry, Gene

Top concepts (fields/topics) attached by OpenAlex

Cited by

1

Total citation count in OpenAlex

Citations by year (recent)

2025: 1

Per-year citation counts (last 5 years)

Related works (count)

10

Other works algorithmically related by OpenAlex

Full payload