Adversarial Input

Adversarial input research focuses on developing and mitigating vulnerabilities in machine learning models, particularly large language models (LLMs) and deep neural networks (DNNs), by crafting inputs designed to elicit incorrect or harmful outputs. Current research emphasizes developing novel attack methods, such as prompt injection and image manipulation techniques, alongside robust defenses including adversarial training, invariance regularization, and prompt rewriting. This field is crucial for ensuring the safe and reliable deployment of AI systems across various applications, from autonomous vehicles to medical diagnosis, by improving model robustness and trustworthiness.

Papers

May 23, 2022

Learning to Ignore Adversarial Attacks
Yiming Zhang, Yangqiaoyu Zhou, Samuel Carton, Chenhao Tan
LeArning Abstract Adversarial Attack Adversarial Example Natural Language Processing Model Adversarial Input Token Level Attack

May 21, 2022

Gradient Concealment: Free Lunch for Defending Adversarial Attacks
Sen Pei, Jiaxi Sun, Xiaopeng Zhang, Gaofeng Meng
Deep Learning Adversarial Attack Adversarial Example Adversarial Input Free Lunch Gradient Obfuscation

April 21, 2022

A Mask-Based Adversarial Defense Scheme
Weizhen Xu, Chenyi Zhang, Fangzhen Zhao, Liangda Fang
Adversarial Attack Adversarial Input Adversarial Defense Adversarial Mask

April 18, 2022

Sardino: Ultra-Fast Dynamic Ensemble for Secure Visual Sensing at Mobile Edge
Qun Song, Zhenyu Yan, Wenjie Luo, Rui Tan
Adversarial Example Adversarial Input Visual Sensor Mobile Edge Adaptive Adversary

April 14, 2022

Automatic Fake News Detection: Are current models "fact-checking" or "gut-checking"?
Ian Kelk, Benjamin Basseri, Wee Yi Lee, Richard Qiu, Chris Tanner
Full Model Fact Checking Fake News Detection User Sentiment Adversarial Input Fake News Detection Model Emotion Vector

April 13, 2022

Synthesizing Adversarial Visual Scenarios for Model-Based Robotic Control
Shubhankar Agarwal, Sandeep P. Chinchali
Adversarial Example Model Predictive Control Model Based Adversarial Input Robust Perception

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

March 29, 2022

Can NMT Understand Me? Towards Perturbation-based Evaluation of NMT Models for Code Generation
Pietro Liguori, Cristina Improta, Simona De Vivo, Roberto Natella, Bojan Cukic, Domenico Cotroneo
Native Robustness Code Generation Neural Machine Translation Adversarial Input Robustness Evaluation Neural Machine Translation Model

March 23, 2022

Input-specific Attention Subnetworks for Adversarial Detection
Emil Biju, Anirudh Sriram, Pratyush Kumar, Mitesh M Khapra
Adversarial Attack Adversarial Example Adversarial Input Adversarial Detection

February 18, 2022

Rethinking Machine Learning Robustness via its Link with the Out-of-Distribution Problem
Abderrahmen Amich, Birhanu Eshete
Native Robustness Adversarial Example Adversarial Input Tiltable Link Robust Machine Learning Machine Learning Robustness

February 12, 2022

Excitement Surfeited Turns to Errors: Deep Learning Testing Framework Based on Excitable Neurons
Haibo Jin, Ruoxi Chen, Haibin Zheng, Jinyin Chen, Yao Cheng, Yue Yu, Xianglong Liu
Deep Neural Network Single Neuron Level Error Feedback Adversarial Input Deep Image Deep Learning Testing Neuron Coverage

January 8, 2022

ADI: Adversarial Dominating Inputs in Vertical Federated Learning Systems
Qi Pang, Yuanyuan Yuan, Shuai Wang, Wenting Zheng
Privacy Preserving Vertical Federated Learning Adversarial Input Adaptive Adversary Joint Inference

December 23, 2021

Adaptive Modeling Against Adversarial Attacks
Zhiwen Yan, Teck Khim Ng
Adversarial Attack Adversarial Training Adversarial Input Adversarial Data Adversarial Defense Adaptive Model Gradient Attack

November 25, 2021

ExPLoit: Extracting Private Labels in Split Learning
Sanjay Kariyappa, Moinuddin K Qureshi
Split Learning Adversarial Input Private Label Label Leakage

November 19, 2021

TnT Attacks! Universal Naturalistic Adversarial Patches Against Deep Neural Network Systems
Bao Gia Doan, Minhui Xue, Shiqing Ma, Ehsan Abbasnejad, Damith C. Ranasinghe
Generative Adversarial Network Adversarial Example Adversarial Patch Adversarial Input Deep Learning System Adversarial Patch Attack Energetic Material

November 17, 2021

Two-Face: Adversarial Audit of Commercial Face Recognition Systems
Siddharth D Jaiswal, Karthikeya Duggirala, Abhisek Dash, Animesh Mukherjee
Absolute Stance Bias Adversarial Input Recognition System

November 16, 2021

Improving the robustness and accuracy of biomedical language models through adversarial training
Milad Moradi, Matthias Samwald
Language Model Native Robustness Adversarial Training Adversarial Input Biomedical Natural Language Processing Clinical Natural Language Processing