Adversarial Trigger

Adversarial triggers are carefully crafted inputs designed to manipulate the behavior of machine learning models, particularly large language models (LLMs) and neural networks used in image classification and other tasks. Current research focuses on developing increasingly sophisticated methods to generate these triggers, often employing reinforcement learning or gradient-based optimization techniques, and exploring their transferability across different model architectures like BERT, GPT, and various CNN/RNN structures. The ability to create effective adversarial triggers highlights significant vulnerabilities in these models, impacting their reliability and safety in real-world applications, and driving the development of robust defenses.

Papers

November 4, 2024

Attacking Vision-Language Computer Agents via Pop-ups
Yanzhe Zhang, Tao Yu, Diyi Yang
Vision Language Model Autonomous Agent Agent System Accidental Click Adversarial Trigger Sentence Level Attack

October 14, 2024

Out-of-Bounding-Box Triggers: A Stealthy Approach to Cheat Object Detectors
Tao Lin, Lijia Yu, Gaojie Jin, Renjue Li, Peng Wu, Lijun Zhang
Adversarial Robustness Adversarial Patch Bounding Box State of the Art Detector Digital STEALTH Metric Cheating Detection Adversarial Trigger Trigger Point

September 5, 2024

Bypassing DARCY Defense: Indistinguishable Universal Adversarial Triggers
Zuquan Peng, Yuanyuan He, Jianbing Ni, Ben Niu
Adversarial Example Text Generation Adversarial DEfense Adversarial Trigger

August 5, 2024

Can Reinforcement Learning Unlock the Hidden Dangers in Aligned Large Language Models?
Mohammad Bahrami Karkevandi, Nishant Vishwamitra, Peyman Najafirad
Language Model BERTScore Metric Adversarial Trigger Hidden Risk

May 30, 2024

Phantom: General Trigger Attacks on Retrieval Augmented Language Generation
Harsh Chaudhari, Giorgio Severi, John Abascal, Matthew Jagielski, Christopher A. Choquette-Choo, Milad Nasr, Cristina Nita-Rotaru, Alina Oprea
Adversarial Attack Retrieval Augmented Generation Adversarial Pattern Phantom Threat Adversarial Trigger Trigger Attack

April 24, 2024

Universal Adversarial Triggers Are Not Universal
Nicholas Meade, Arkil Patel, Siva Reddy
Language Model Native Robustness Token Sequence Adversarial Trigger

February 21, 2024

Learning to Poison Large Language Models During Instruction Tuning
Yao Qiang, Xiangyu Zhou, Saleh Zare Zade, Mohammad Amin Roshani, Prashant Khanduri, Douglas Zytko, Dongxiao Zhu
Language Model LeArning Abstract Instruction Tuning Data Poisoning Attack Backdoor Trigger Adversarial Trigger

September 1, 2023

Why do universal adversarial attacks work on large language models?: Geometry might be the answer
Varshini Subhash, Anna Bialas, Weiwei Pan, Finale Doshi-Velez
Large Language Model Adversarial Attack Geometric Analysis Top Two Answer Universal Adversarial Adversarial Trigger Adversarial Region

November 28, 2022

Attack on Unfair ToS Clause Detection: A Case Study using Universal Adversarial Triggers
Shanshan Xu, Irina Broda, Rashid Haddad, Marco Negrini, Matthias Grabmair
Natural Language Processing Adversarial Attack Case Study Adversarial Trigger

November 18, 2022

Adversarial Stimuli: Attacking Brain-Computer Interfaces via Perturbed Sensory Events
Bibek Upadhayay, Vahid Behzadan
Brain Computer Interface Motor Imagery Adversarial Trigger

September 14, 2022

Order-Disorder: Imitation Adversarial Attacks for Black-box Neural Ranking Models
Jiawei Liu, Yangyang Kang, Di Tang, Kaisong Song, Changlong Sun, Xiaofeng Wang, Wei Lu, Xiaozhong Liu
Adversarial Attack Adversarial Vulnerability Neural Ranker Adversarial Imitation Adversarial Trigger

May 28, 2022

Contributor-Aware Defenses Against Adversarial Backdoor Attacks
Glenn Dawson, Muhammad Umer, Robi Polikar
Adversarial Attack Multi Step Adversarial Attack Adversarial Backdoor Adversarial Trigger Certified Defense

April 11, 2022

Exploring the Universal Vulnerability of Prompt-based Learning Paradigm
Lei Xu, Yangyi Chen, Ganqu Cui, Hongcheng Gao, Zhiyuan Liu
Pre Trained Language Model Prompt Based Learning Inherent Vulnerability Adversarial Trigger

November 22, 2021

NTD: Non-Transferability Enabled Backdoor Detection
Yinshan Li, Hua Ma, Zhi Zhang, Yansong Gao, Alsharif Abuadbba, Anmin Fu, Yifeng Zheng, Said F. Al-Sarawi, Derek Abbott
Backdoor Attack Task Transferability Backdoor Detection Adversarial Trigger Transferable Backdoor Attack

November 18, 2021

Attacking Deep Learning AI Hardware with Universal Adversarial Perturbation
Mehdi Sadi, B. M. S. Bahar Talukder, Kaniz Mishty, Md Tauhidur Rahman
Adversarial Noise Deep Learning Application Additive Noise Adversarial Trigger