Backdoor Attack

Backdoor attacks exploit vulnerabilities in machine learning models by embedding hidden triggers during training, causing the model to produce malicious outputs when the trigger is present. Current research focuses on developing and mitigating these attacks across various model architectures, including deep neural networks, vision transformers, graph neural networks, large language models, and spiking neural networks, with a particular emphasis on understanding attack mechanisms and developing robust defenses in federated learning and generative models. The significance of this research lies in ensuring the trustworthiness and security of increasingly prevalent machine learning systems across diverse applications, ranging from object detection and medical imaging to natural language processing and autonomous systems.

Papers

January 14, 2025

Energy Backdoor Attack to Deep Neural Networks
Hanene F. Z. Brachemi Meftah, Wassim Hamidouche, Sid Ahmed Fezza, Olivier Déforges, Kassem Kallas
Deep Neural Network Backdoor Attack Inference Time

January 13, 2025

A4O: All Trigger for One sample
Duc Anh Vu, Anh Tuan Tran, Cong Tran, Cuong Pham
Backdoor Attack External Sample Backdoor Trigger Natural Trigger

January 11, 2025

SafeSplit: A Novel Defense Against Client-Side Backdoor Attacks in Split Learning
Phillip Rieger, Alessandro Pegoraro, Kavita Kumari, Tigist Abera, Jonathan Knauer, Ahmad-Reza Sadeghi
LeArning Abstract Backdoor Attack Split Learning Backdoor Defense Split and Fit Defense Method

January 10, 2025

Fine-tuning is Not Fine: Mitigating Backdoor Attacks in GNNs with Limited Clean Data
Jiale Zhang, Bosen Rao, Chengcheng Zhu, Xiaobing Sun, Qingming Li, Haibo Hu, Xiapu Luo, Qingqing Ye, Shouling Ji
Graph Neural Network Fine Tuning Backdoor Attack Graph Attention Limited Data Backdoor Defense Neural Backdoor

January 6, 2025

MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation
Yidong Ding, Jiafei Niu, Ping Yi
Language Model Adversarial Training Backdoor Attack NLP Model Attention Distillation

January 5, 2025

Backdoor Token Unlearning: Exposing and Defending Backdoors in Pretrained Language Models
Peihai Jiang, Xixiang Lyu, Yige Li, Jing Ma
Backdoor Attack Pretrained Language Model Backdoor Learning Backdoor Behavior

January 4, 2025

BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors
Chia-Yi Hsu, Yu-Lin Tsai, Yu Zhe, Yan-Lun Chen, Chih-Hsun Lin, Chia-Mu Yu, Yang Zhang, Chun-Ying Huang, Jun Sakuma
Backdoor Attack Attack Success Rate Task Vector Multimodal Backdoor Faithful Model

January 3, 2025

January 2, 2025

December 29, 2024

December 26, 2024

CL-attack: Textual Backdoor Attacks via Cross-Lingual Triggers
Jingyi Zheng, Tianyi Hu, Tianshuo Cong, Xinlei He
Backdoor Attack Textual Backdoor Attack Trigger Word Cross Lingual Backdoor Attack

December 25, 2024

Injecting Bias into Text Classification Models using Backdoor Attacks
A. Dilara Yavuz, M. Emre Gursoy
Backdoor Attack Text Classification Bias Evaluation Text Classification Model Label Attack Bias Challenge

December 23, 2024

December 22, 2024

A Backdoor Attack Scheme with Invisible Triggers Based on Model Architecture Modification
Yuan Ma, Xu Ma, Jiankang Wei, Jinmeng Tang, Xiaoyu Zhang, Yilun Lyu, Kehao Chen, Jingtong Huang
Full Model Backdoor Attack Imperceptible Trigger

December 21, 2024

TrojFlow: Flow Models are Natural Targets for Trojan Attacks
Zhengyang Qi, Xiaohua Xu
Generative Model Backdoor Attack Multiple Target Trojan Attack Flow Based Model

December 17, 2024

BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection
He Cheng, Depeng Xu, Shuhan Yuan
Anomaly Detection Backdoor Attack Semi Supervised Image Anomaly Detection Clean Label Backdoor Attack

Backdoor Attack

Papers

Energy Backdoor Attack to Deep Neural Networks

A4O: All Trigger for One sample

SafeSplit: A Novel Defense Against Client-Side Backdoor Attacks in Split Learning

Fine-tuning is Not Fine: Mitigating Backdoor Attacks in GNNs with Limited Clean Data

MBTSAD: Mitigating Backdoors in Language Models Based on Token Splitting and Attention Distillation

Backdoor Token Unlearning: Exposing and Defending Backdoors in Pretrained Language Models

BADTV: Unveiling Backdoor Threats in Third-Party Task Vectors

Mingling with the Good to Backdoor Federated Learning

BLAST: A Stealthy Backdoor Leverage Attack against Cooperative Multi-Agent Deep Reinforcement Learning based Systems

Stealthy Backdoor Attack to Real-world Models in Android Apps

HoneypotNet: Backdoor Attacks Against Model Extraction

Cut the Deadwood Out: Post-Training Model Purification with Selective Module Substitution

Defending Multimodal Backdoored Models by Repulsive Visual Prompt Tuning

CL-attack: Textual Backdoor Attacks via Cross-Lingual Triggers

Injecting Bias into Text Classification Models using Backdoor Attacks

Trading Devil RL: Backdoor attack via Stock market, Bayesian Optimization and Reinforcement Learning

Double Landmines: Invisible Textual Backdoor Attacks based on Dual-Trigger

A Backdoor Attack Scheme with Invisible Triggers Based on Model Architecture Modification

TrojFlow: Flow Models are Natural Targets for Trojan Attacks

BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection