Adversarial Sample

Adversarial samples are inputs designed to intentionally mislead machine learning models, primarily by introducing small, imperceptible perturbations to otherwise correctly classified data. Current research focuses on developing more robust models through techniques like adversarial training, purification methods using generative models (e.g., GANs), and exploring the vulnerabilities of various architectures, including convolutional neural networks, recurrent networks, and large language models. Understanding and mitigating the impact of adversarial samples is crucial for ensuring the reliability and security of machine learning systems across diverse applications, from cybersecurity to medical diagnosis.

Papers

May 10, 2023

A Black-Box Attack on Code Models via Representation Nearest Neighbor Search
Jie Zhang, Wei Ma, Qiang Hu, Shangqing Liu, Xiaofei Xie, Yves Le Traon, Yang Liu
Adversarial Example Adversarial Training Adversarial Sample Black Box Attack Code Model

May 5, 2023

White-Box Multi-Objective Adversarial Attack on Dialogue Generation
Yufei Li, Zexin Li, Yingfan Gao, Cong Liu
Adversarial Sample Pre Trained Transformer Dialogue Generation White Box Attack Performance

March 30, 2023

Generating Adversarial Samples in Mini-Batches May Be Detrimental To Adversarial Robustness
Timothy Redgrave, Colton Crum
Adversarial Attack Adversarial Robustness Adversarial Sample Mini Batch Adversarial Data Collection

March 29, 2023

Beyond Empirical Risk Minimization: Local Structure Preserving Regularization for Improving Adversarial Robustness
Wei Wei, Jiahuan Zhou, Ying Wu
Adversarial Training Adversarial Robustness Empirical Risk Minimization Adversarial Sample

March 28, 2023

March 27, 2023

EMShepherd: Detecting Adversarial Samples via Side-channel Leakage
Ruyi Ding, Cheng Gongye, Siyue Wang, Aidong Ding, Yunsi Fei
Adversarial Attack Adversarial Sample Side Channel Adversarial Detection Wave Propagation

March 22, 2023

March 19, 2023

FVQA 2.0: Introducing Adversarial Samples into Fact-based Visual Question Answering
Weizhe Lin, Zhilin Wang, Bill Byrne
Adversarial Sample Knowledge Based Visual Question Answering

March 10, 2023

March 7, 2023

Logit Margin Matters: Improving Transferable Targeted Adversarial Attack by Logit Calibration
Juanjuan Weng, Zhiming Luo, Zhun Zhong, Shaozi Li, Nicu Sebe
Adversarial Example Loss Function Cross Entropy Loss Adversarial Sample Transferable Adversarial Attack Logit Calibration Adaptive Margin

March 6, 2023

Adversarial Sampling for Fairness Testing in Deep Neural Network
Tosin Ige, William Marfo, Justin Tonkinson, Sikiru Adewale, Bolanle Hafiz Matti
Deep Neural Network Adversarial Attack Adversarial Sample Adversarial Data Collection Adversarial Training Algorithm FAIrness Testing

February 11, 2023

Mutation-Based Adversarial Attacks on Neural Text Detectors
Gongbo Liang, Jesus Guerrero, Izzat Alsmadi
Adversarial Attack Adversarial Sample Text Detection Text Detector White Box Adversarial Attack

January 31, 2023

Inference Time Evidences of Adversarial Attacks for Forensic on Transformers
Hugo Lemarchant, Liangzi Li, Yiming Qian, Yuta Nakashima, Hajime Nagahara
Adversarial Attack Vision Transformer Transformer Megatron Decepticons Adversarial Perturbation Adversarial Sample Latent Vector Forensic Analysis

December 11, 2022

General Adversarial Defense Against Black-box Attacks via Pixel Level and Feature Level Distribution Alignments
Xiaogang Xu, Hengshuang Zhao, Philip Torr, Jiaya Jia
Adversarial Attack Adversarial Sample Distribution Alignment Deep Generative Self Supervised Pixel

November 29, 2022

A3T: Accuracy Aware Adversarial Training
Enes Altinisik, Safa Messaoud, Husrev Taha Sencar, Sanjay Chawla
Adversarial Training Adversarial Sample Better Generalization Adversarial Evaluation Labeled Sample

November 26, 2022

Supervised Contrastive Prototype Learning: Augmentation Free Robust Neural Network
Iordanis Fostiropoulos, Laurent Itti
Contrastive Learning Adversarial Example Contrastive Loss Adversarial Sample Robust Data

November 20, 2022

Feature Weaken: Vicinal Data Augmentation for Classification
Songhao Jiang, Yan Chu, Tianxing Ma, Tianning Zang
Classification Code Adversarial Sample Novel Training Feature Mixing Non Robust Feature