Self-Supervised Contrastive Learning with Adversarial Perturbations for Defending Word Substitution-based Attacks

2021/07/15 by Zhao Meng, Yihan Dong, Meng, Zhao +5 · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Computation and Language (cs.CL) #FOS: Computer and information sciences #Natural Language Processing Techniques #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2107.07610

openalex publication_date 2021/07/15 · openalex created_date 2022/11/11 · openalex updated_date 2026/07/28

Abstract

In this paper, we present an approach to improve the robustness of BERT language models against word substitution-based adversarial attacks by leveraging adversarial perturbations for self-supervised contrastive learning. We create a word-level adversarial attack generating hard positives on-the-fly as adversarial examples during contrastive learning. In contrast to previous works, our method improves model robustness without using any labeled data. Experimental results show that our method improves robustness of BERT against four different word substitution-based adversarial attacks, and combining our method with adversarial training gives higher robustness than adversarial training alone. As our method improves the robustness of BERT purely with unlabeled data, it opens up the possibility of using large text datasets to train robust language models against word substitution-based adversarial attacks.

Cited by

• Unsupervised Candidate Ranking for Lexical Substitution via Holistic Sentence Semantics

Related

• SCAT: Robust Self-supervised Contrastive Learning via Adversarial Training for Text Classification
• Fast Adversarial Training against Textual Adversarial Attacks
• BERT is Robust! A Case Against Synonym-Based Adversarial Examples in Text Classification
• BERT-ATTACK: Adversarial Attack Against BERT Using BERT
• SemRoDe: Macro Adversarial Training to Learn Representations That are Robust to Word-Level Attacks