Holistic Adversarial Robustness of Deep Learning Models

TL;DR

This paper offers a comprehensive overview of adversarial robustness in deep learning, covering attacks, defenses, verification, and applications to enhance safety and reliability.

Contribution

It provides a holistic survey of research topics and foundational principles in adversarial robustness for deep learning models.

Findings

Summarizes key attack and defense strategies.

Highlights verification methods for robustness.

Discusses emerging applications and challenges.

Abstract

Adversarial robustness studies the worst-case performance of a machine learning model to ensure safety and reliability. With the proliferation of deep-learning-based technology, the potential risks associated with model development and deployment can be amplified and become dreadful vulnerabilities. This paper provides a comprehensive overview of research topics and foundational principles of research methods for adversarial robustness of deep learning models, including attacks, defenses, verification, and novel applications.