Skip to main navigation Skip to search Skip to main content

Semantic-Preserving Adversarial Text Attacks

  • Xinghao Yang
  • , Yongshun Gong
  • , Weifeng Liu
  • , James Bailey
  • , Dacheng Tao
  • , Wei Liu

Research output: Contribution to journal › Article › Research › peer-review

Abstract

Deep learning models are known immensely brittle to adversarial text examples. Existing text adversarial attack strategies can be roughly divided into character-level, word-level, and sentence-level attacks. Despite the success brought by recent text attack methods, how to induce misclassification with minimal text modifications while keeping the lexical correctness, syntactic soundness, and semantic consistency is still a challenge. In this paper, we devise a Bigram and Unigram-based adaptive Semantic Preservation Optimization (BU-SPO) approach which attacks text documents not only at a unigram word level but also at a bigram level to avoid generating meaningless sentences. We also present a hybrid attack strategy that collects substitution words from both synonyms and sememe candidates, to enrich the potential candidate set. Besides, a Semantic Preservation Optimization (SPO) method is devised to determine the word substitution priority and reduce the perturbation cost. Furthermore, we constrain the SPO with a semantic Filter (dubbed SPOF) to improve the semantic similarity. To estimate the effectiveness of our proposed methods, BU-SPO and BU-SPOF, we attack four victim deep learning models trained on three text datasets. Experimental results demonstrate that our approaches accomplish the highest semantics consistency and attack success rates by making minimal word modifications compared with competitive methods.

Original languageEnglish
Pages (from-to)583-595
Number of pages13
JournalIEEE Transactions on Sustainable Computing
Volume8
Issue number4
DOIs
Publication statusPublished - Oct 2023
Externally publishedYes

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

  1. SDG 7 - Affordable and Clean Energy
    SDG 7 Affordable and Clean Energy

Keywords

  • Adversarial machine learning
  • natural language processing
  • text attack

Cite this