Adversarial Example

Adversarial examples are subtly altered inputs designed to fool machine learning models, primarily deep neural networks (DNNs), into making incorrect predictions. Current research focuses on improving model robustness against these attacks, exploring techniques like ensemble methods, multi-objective representation learning, and adversarial training, often applied to architectures such as ResNets and Vision Transformers. Understanding and mitigating the threat of adversarial examples is crucial for ensuring the reliability and security of AI systems across diverse applications, from image classification and natural language processing to malware detection and autonomous driving. The development of robust defenses and effective attack detection methods remains a significant area of ongoing investigation.

Papers

November 18, 2024

Theoretical Corrections and the Leveraging of Reinforcement Learning to Enhance Triangle Attack
Nicole Meng, Caleb Manicke, David Chen, Yingjie Lao, Caiwen Ding, Pengyu Hong, Kaleel Mahmood
Reinforcement Learning Adversarial Example Cyber Attack Supervised Correction

November 15, 2024

A Hard-Label Cryptanalytic Extraction of Non-Fully Connected Deep Neural Networks using Side-Channel Attacks
Benoit Coqueret, Mathieu Carbone, Olivier Sentieys, Gabriel Zaid
Neural Network Deep Neural Network Adversarial Example Network Programming Deep Network Side Channel Hard Label Attack

November 12, 2024

November 11, 2024

Adversarial Detection with a Dynamically Stable System
Xiaowei Long, Jie Lin, Xiangyuan Yang
Adversarial Example Adversarial Detection

November 7, 2024

Neural Fingerprints for Adversarial Attack Detection
Haim Fisher, Moni Shahar, Yehezkel S. Resheff
Deep Learning Model Adversarial Example Image Classification Supervised ImageNet Adversarial Attack Detection

November 5, 2024

Enhancing Adversarial Robustness via Uncertainty-Aware Distributional Adversarial Training
Junhao Dong, Xinghua Qu, Z. Jane Wang, Yew-Soon Ong
Adversarial Example Adversarial Training Adversarial Robustness Domain Adversarial

November 4, 2024

November 1, 2024

Replace-then-Perturb: Targeted Adversarial Attacks With Visual Reasoning for Vision-Language Models
Jonggyu Jang, Hyeonsu Lyu, Jungyeon Koh, Hyun Jong Yang
Adversarial Attack Vision Language Model Adversarial Example Visual Language Model Visual Reasoning Additive Perturbation

October 31, 2024

October 29, 2024

October 28, 2024

Attacking Misinformation Detection Using Adversarial Examples Generated by Language Models
Piotr Przybyła
Large Language Model Language Model Adversarial Example Text Classification Unreliable Source

October 26, 2024

Transferable Adversarial Attacks on SAM and Its Downstream Models
Song Xia, Wenhan Yang, Yi Yu, Xun Lin, Henghui Ding, Lingyu Duan, Xudong Jiang
Adversarial Attack Adversarial Example Segment Anything Model Transferable Adversarial Attack Adversarial Risk Downstream Model Transfer Based Adversarial Attack

Adversarial Example

Papers

Theoretical Corrections and the Leveraging of Reinforcement Learning to Enhance Triangle Attack

A Hard-Label Cryptanalytic Extraction of Non-Fully Connected Deep Neural Networks using Side-Channel Attacks

IAE: Irony-based Adversarial Examples for Sentiment Analysis Systems

Chain Association-based Attacking and Shielding Natural Language Processing Systems

Adversarial Detection with a Dynamically Stable System

Neural Fingerprints for Adversarial Attack Detection

Enhancing Adversarial Robustness via Uncertainty-Aware Distributional Adversarial Training

Semantic-Aligned Adversarial Evolution Triangle for High-Transferability Vision-Language Attack

LiDAttack: Robust Black-box Attack on LiDAR-based Object Detection

Replace-then-Perturb: Targeted Adversarial Attacks With Visual Reasoning for Vision-Language Models

Protecting Feed-Forward Networks from Adversarial Attacks Using Predictive Coding

Noise as a Double-Edged Sword: Reinforcement Learning Exploits Randomized Defenses in Neural Networks

Wide Two-Layer Networks can Learn from Adversarial Perturbations

Teaching a Language Model to Distinguish Between Similar Details using a Small Adversarial Training Set

CausalDiff: Causality-Inspired Disentanglement via Diffusion Model for Adversarial Defense

One Prompt to Verify Your Models: Black-Box Text-to-Image Models Verification via Non-Transferable Adversarial Attacks

CausAdv: A Causal-based Framework for Detecting Adversarial Examples

On the Robustness of Adversarial Training Against Uncertainty Attacks

Attacking Misinformation Detection Using Adversarial Examples Generated by Language Models

Transferable Adversarial Attacks on SAM and Its Downstream Models