CT-GAT: Cross-Task Generative Adversarial Attack based on Transferability

Minxuan Lv, Chengwei Dai, Kun Li, Wei Zhou, Songlin Hu

Submission Type: Regular Long Paper

Submission Track: Ethics in NLP

Submission Track 2: Natural Language Generation

Keywords: Adversarial Attacks, Transferability, Generative Methods

Abstract: Neural network models are vulnerable to adversarial examples, and adversarial transferability further increases the risk of adversarial attacks. Current methods based on transferability often rely on substitute models, which can be impractical and costly in real-world scenarios due to the unavailability of training data and the victim model's structural details. In this paper, we propose a novel approach that directly constructs adversarial examples by extracting transferable features across various tasks. Our key insight is that adversarial transferability can extend across different tasks. Specifically, we train a sequence-to-sequence generative model named CT-GAT (Cross-Task Generative Adversarial Attack) using adversarial sample data collected from multiple tasks to acquire universal adversarial features and generate adversarial examples for different tasks.We conduct experiments on ten distinct datasets, and the results demonstrate that our method achieves superior attack performance with small cost.

Submission Number: 565