Adversarial Attacks and Defenses in Images, Graphs and Text: A Review

TL;DR

This survey comprehensively reviews recent adversarial attack and defense techniques across images, graphs, and text domains, highlighting current threats and countermeasures in deep neural networks.

Contribution

It provides a systematic overview of state-of-the-art algorithms for generating adversarial examples and their defenses across multiple data types.

Findings

Adversarial attacks pose significant threats to DNNs in various data domains.

Countermeasures have been developed with varying success across data types.

The survey identifies key challenges and future directions in adversarial robustness.

Abstract

Deep neural networks (DNN) have achieved unprecedented success in numerous machine learning tasks in various domains. However, the existence of adversarial examples has raised concerns about applying deep learning to safety-critical applications. As a result, we have witnessed increasing interests in studying attack and defense mechanisms for DNN models on different data types, such as images, graphs and text. Thus, it is necessary to provide a systematic and comprehensive overview of the main threats of attacks and the success of corresponding countermeasures. In this survey, we review the state of the art algorithms for generating adversarial examples and the countermeasures against adversarial examples, for the three popular data types, i.e., images, graphs and text.