Perturbation Attack

Perturbation attacks involve subtly altering inputs to machine learning models, particularly deep neural networks, to cause misclassification or inaccurate outputs. Current research focuses on developing increasingly effective attack strategies, including those targeting specific model architectures (e.g., vision foundation models, no-reference image quality metrics) and employing diverse perturbation methods (e.g., pixel-wise adjustments under various norms, universal perturbations). This research is crucial for evaluating and improving the robustness of machine learning systems across various applications, from autonomous driving to medical image analysis, where the reliability of model predictions is paramount. Understanding and mitigating these attacks is essential for ensuring the safe and trustworthy deployment of AI systems.

Papers

October 13, 2024

Understanding Robustness of Parameter-Efficient Tuning for Image Classification
Jiacheng Ruan, Xian Gao, Suncheng Xiang, Mingye Xie, Ting Liu, Yuzhuo Fu
Native Robustness Classification Code Adversarial Sample Parameter Efficient Tuning Learnable Parameter Perturbation Attack Natural Robustness Research

July 18, 2024

PG-Attack: A Precision-Guided Adversarial Attack Framework Against Vision Foundation Models for Autonomous Driving
Jiyuan Fu, Zhaoyu Chen, Kaixun Jiang, Haijing Guo, Shuyong Gao, Wenqiang Zhang
Adversarial Attack Autonomous Driving Vision Foundation Model Black Box Adversarial Attack Perturbation Attack

May 24, 2023

Fast Adversarial CNN-based Perturbation Attack on No-Reference Image- and Video-Quality Metrics
Ekaterina Shumitskaya, Anastasia Antsiferova, Dmitriy Vatolin
Adversarial Perturbation Reference Image Fast Adversarial Training Traditional Adversarial Video Quality Metric Reference Quality Assessment Perturbation Attack Reference Quality

April 16, 2023

A Random-patch based Defense Strategy Against Physical Attacks for Face Recognition Systems
JiaHao Xie, Ye Luo, Jianwei Lu
Adaptive Attack Complex Deep Face Recognition System Perturbation Attack Physical Attack

December 5, 2022

Multiple Perturbation Attack: Attack Pixelwise Under Different $\ell_p$-norms For Better Adversarial Performance
Ngoc N. Tran, Anh Tuan Bui, Dinh Phung, Trung Le
Adversarial Attack Adversarial Example \Ell_p$ Norm Perturbation Attack Better Adversarial Robustness ImageNet Attack

November 9, 2022

Optimized Global Perturbation Attacks For Brain Tumour ROI Extraction From Binary Classification Models
Sajith Rajapaksa, Farzad Khalvati
Tumor Segmentation Brain Tumor Segmentation Binary Classifier 3D Brain Tumor Binary Label Perturbation Attack

November 1, 2022

Universal Perturbation Attack on Differentiable No-Reference Image- and Video-Quality Metrics
Ekaterina Shumitskaya, Anastasia Antsiferova, Dmitriy Vatolin
Reference Image Video Quality Metric Reference Quality Assessment Perturbation Attack Universal Perturbation

October 2, 2022

Adaptive Smoothness-weighted Adversarial Training for Multiple Perturbations with Its Stability Analysis
Jiancong Xiao, Zeyu Qin, Yanbo Fan, Baoyuan Wu, Jue Wang, Zhi-Quan Luo
Adversarial Example Adversarial Training Adversarial Robustness Adaptive Importance Adversarial Loss Additive Perturbation Stability Analysis Perturbation Attack

September 24, 2022

Strong Transferable Adversarial Attacks via Ensembled Asymptotically Normal Distribution Learning
Zhengwei Fang, Rui Wang, Tao Huang, Liping Jing
Adversarial Example Adversarial Perturbation Transferable Adversarial Attack Perturbation Attack Normality Learning

August 12, 2022

Unifying Gradients to Improve Real-world Robustness for Deep Networks
Yingwen Wu, Sizhe Chen, Kun Fang, Xiaolin Huang
Adversarial Attack Deep Network QuEry Based Attack Perturbation Attack

March 20, 2022

Adversarial Parameter Attack on Deep Neural Networks
Lijia Yu, Yihan Wang, Xiao-Shan Gao
Deep Neural Network Perturbation Attack Adversarial Parameter

Perturbation Attack

Papers

Understanding Robustness of Parameter-Efficient Tuning for Image Classification

PG-Attack: A Precision-Guided Adversarial Attack Framework Against Vision Foundation Models for Autonomous Driving

Fast Adversarial CNN-based Perturbation Attack on No-Reference Image- and Video-Quality Metrics

A Random-patch based Defense Strategy Against Physical Attacks for Face Recognition Systems

Multiple Perturbation Attack: Attack Pixelwise Under Different $\ell_p$-norms For Better Adversarial Performance

Optimized Global Perturbation Attacks For Brain Tumour ROI Extraction From Binary Classification Models

Universal Perturbation Attack on Differentiable No-Reference Image- and Video-Quality Metrics

Adaptive Smoothness-weighted Adversarial Training for Multiple Perturbations with Its Stability Analysis

Strong Transferable Adversarial Attacks via Ensembled Asymptotically Normal Distribution Learning

Unifying Gradients to Improve Real-world Robustness for Deep Networks

Adversarial Parameter Attack on Deep Neural Networks