Textual Adversarial Example

Textual adversarial examples are subtly altered text inputs designed to deceive natural language processing (NLP) models, highlighting vulnerabilities in their robustness. Current research focuses on developing more effective attack methods, often employing techniques like synonym substitution and phrase-level manipulation within various model architectures, including BERT and other large language models (LLMs), and exploring defenses such as test-time adaptation and manifold-based approaches. Understanding and mitigating these vulnerabilities is crucial for ensuring the reliability and security of NLP systems in real-world applications, particularly in safety-critical domains.

Papers

August 15, 2024

Evaluating Text Classification Robustness to Part-of-Speech Adversarial Examples
Anahita Samadi, Allison Sullivan
Adversarial Example CNN Model Linguistic Capability Textual Adversarial Example Robust Text Classification

May 24, 2024

Large Language Model Sentinel: LLM Agent for Adversarial Purification
Guang Lin, Qibin Zhao
Adversarial Attack Adversarial Robustness LLM Agent Adversarial DEfense Adversarial Behavior Textual Adversarial Example

February 26, 2024

RoCoIns: Enhancing Robustness of Large Language Models through Code-Style Instructions
Yuansen Zhang, Xiao Wang, Zhiheng Xi, Han Xia, Tao Gui, Qi Zhang, Xuanjing Huang
Native Robustness Real World Code Adversarial Sample Natural Language Instruction Textual Adversarial Example Adversarial Context

February 5, 2024

Arabic Synonym BERT-based Adversarial Examples for Text Classification
Norah Alshahrani, Saied Alshahrani, Esma Wali, Jeanna Matthews
Adversarial Attack Adversarial Example Text Classification Textual Adversarial Attack Textual Adversarial Example Arabic BERt

February 2, 2024

HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Text
Han Liu, Zhi Xu, Xiaotong Zhang, Feng Zhang, Fenglong Ma, Hongyang Chen, Hong Yu, Xianchao Zhang
Adversarial Example Text Modality New Attack Textual Adversarial Example Hard Label Attack

October 29, 2023

Robustifying Language Models with Test-Time Adaptation
Noah Thomas McDermott, Junfeng Yang, Chengzhi Mao
Language Model Test Time Adaptation Large Scale Language Model Adversarial Corruption Textual Adversarial Example

July 3, 2023

Interpretability and Transparency-Driven Detection and Transformation of Textual Adversarial Examples (IT-DT)
Bushra Sabir, M. Ali Babar, Sharif Abuadbba
Adversarial Attack Adversarial Example Inherent Interpretability Generative Adversarial Adversarial Input Character Transformation Textual Adversarial Example

May 6, 2023

The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples
Heng Yang, Ke Li
Adversarial Attack Adversarial Example Semantics Surfaced Adversarial Scenario Adversarial Defense Textual Adversarial Example

April 18, 2023

Masked Language Model Based Textual Adversarial Example Detection
Xiaomei Zhang, Zhaoxi Zhang, Qi Zhong, Xufei Zheng, Yanjun Zhang, Shengshan Hu, Leo Yu Zhang
Adversarial Attack Adversarial Example Masked Language Adversarial DEfense Textual Adversarial Example

March 9, 2023

BeamAttack: Generating High-quality Textual Adversarial Examples through Beam Search and Mixed Semantic Spaces
Hai Zhu, Qingyang Zhao, Yuren Wu
Adversarial Example Adversarial Training Beam Search Semantic Space Textual Adversarial Example Word Level Attack Textual Attack

March 1, 2023

Frauds Bargain Attack: Generating Adversarial Text Samples via Word Manipulation Process
Mingze Ni, Zhensu Sun, Wei Liu
Natural Language Processing Adversarial Example Textual Adversarial Example Procurement Fraud

November 5, 2022

Textual Manifold-based Defense Against Natural Language Adversarial Examples
Dang Minh Nguyen, Luu Anh Tuan
Adversarial Attack Adversarial Example Adversarial Text Adversarial Image Textual Adversarial Example Manifold Attack

October 19, 2022

Why Should Adversarial Perturbations be Imperceptible? Rethink the Research Paradigm in Adversarial NLP
Yangyi Chen, Hongcheng Gao, Ganqu Cui, Fanchao Qi, Longtao Huang, Zhiyuan Liu, Maosong Sun
Adversarial Perturbation Adversarial Sample NLP Research Adversarial Objective Textual Adversarial Example Research Paradigm Adversarial Natural Language

July 21, 2022

Rethinking Textual Adversarial Defense for Pre-trained Language Models
Jiayi Wang, Rongzhou Bao, Zhuosheng Zhang, Hai Zhao
Adversarial Attack Adversarial Example Pre Trained Language Model Textual Adversarial Attack Textual Adversarial Example

May 22, 2022

Phrase-level Textual Adversarial Attack with Label Preservation
Yibin Lei, Yu Cao, Dianqi Li, Tianyi Zhou, Meng Fang, Mykola Pechenizkiy
Adversarial Sample Latent Vulnerability Level Perturbation Textual Adversarial Example

April 29, 2022

Detecting Textual Adversarial Examples Based on Distributional Characteristics of Data Representations
Na Liu, Mark Dras, Wei Emma Zhang
Adversarial Attack Adversarial Example Adversarial Training Data Representation Textual Adversarial Example Distributional Property Sentence Level Attack

April 10, 2022

"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attacks
Edoardo Mosca, Shreyash Agarwal, Javier Rando, Georg Groh
Natural Language Processing Adversarial Attack Adversarial Input Textual Adversarial Example Word Level Attack Suspicious Commenter Adversarial Natural Language

February 17, 2022

Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition
Chao-Han Huck Yang, Zeeshan Ahmed, Yile Gu, Joseph Szurley, Roger Ren, Linda Liu, Andreas Stolcke, Ivan Bulyko
Adversarial Example End to End Adversarial Noise Textual Adversarial Example Bayesian Neural

November 4, 2021

Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models
Boxin Wang, Chejian Xu, Shuohang Wang, Zhe Gan, Yu Cheng, Jianfeng Gao, Ahmed Hassan Awadallah, Bo Li
Language Model Adversarial Attack Robustness Evaluation Multi Task Benchmark Textual Adversarial Attack Textual Adversarial Example

Textual Adversarial Example

Papers

Evaluating Text Classification Robustness to Part-of-Speech Adversarial Examples

Large Language Model Sentinel: LLM Agent for Adversarial Purification

RoCoIns: Enhancing Robustness of Large Language Models through Code-Style Instructions

Arabic Synonym BERT-based Adversarial Examples for Text Classification

HQA-Attack: Toward High Quality Black-Box Hard-Label Adversarial Attack on Text

Robustifying Language Models with Test-Time Adaptation

Interpretability and Transparency-Driven Detection and Transformation of Textual Adversarial Examples (IT-DT)

The Best Defense is Attack: Repairing Semantics in Textual Adversarial Examples

Masked Language Model Based Textual Adversarial Example Detection

BeamAttack: Generating High-quality Textual Adversarial Examples through Beam Search and Mixed Semantic Spaces

Frauds Bargain Attack: Generating Adversarial Text Samples via Word Manipulation Process

Textual Manifold-based Defense Against Natural Language Adversarial Examples

Why Should Adversarial Perturbations be Imperceptible? Rethink the Research Paradigm in Adversarial NLP

Rethinking Textual Adversarial Defense for Pre-trained Language Models

Phrase-level Textual Adversarial Attack with Label Preservation

Detecting Textual Adversarial Examples Based on Distributional Characteristics of Data Representations

"That Is a Suspicious Reaction!": Interpreting Logits Variation to Detect NLP Adversarial Attacks

Mitigating Closed-model Adversarial Examples with Bayesian Neural Modeling for Enhanced End-to-End Speech Recognition

Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models