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

June 16, 2024

Imperceptible Rhythm Backdoor Attacks: Exploring Rhythm Transformation for Embedding Undetectable Vulnerabilities on Speech Recognition
Wenhan Yao, Jiangkun Yang, Yongqiang He, Jia Liu, Weiping Wen
Automatic Speech Recognition Speech Recognition Backdoor Attack Speech Recognition System Software Vulnerability Audio Spectrogram

June 15, 2024

June 14, 2024

Robustness-Inspired Defense Against Backdoor Attacks on Graph Neural Networks
Zhiwei Zhang, Minhua Lin, Junjie Xu, Zongyu Wu, Enyan Dai, Suhang Wang
Graph Neural Network Backdoor Attack Graph Backdoor Attack

June 10, 2024

June 9, 2024

June 5, 2024

June 3, 2024

Unelicitable Backdoors in Language Models via Cryptographic Transformer Circuits
Andis Draguns, Andrew Gritsevskiy, Sumeet Ramesh Motwani, Charlie Rogers-Smith, Jeffrey Ladish, Christian Schroeder de Witt
Language Model Backdoor Attack State of the Art Robustness Encrypted Vision Transformer Backdoor Attacker

June 2, 2024

May 31, 2024

GANcrop: A Contrastive Defense Against Backdoor Attacks in Federated Learning
Xiaoyun Gan, Shanyu Gan, Taizhi Su, Peng Liu
Generative Adversarial Network Backdoor Attack Contrastive Strategy

May 30, 2024

DiffPhysBA: Diffusion-based Physical Backdoor Attack against Person Re-Identification in Real-World
Wenli Sun, Xinyang Jiang, Dongsheng Li, Cairong Zhao
Backdoor Attack Real World Internet Service Domain State of the Art Defense Digital STEALTH Metric Physical Backdoor Attack

May 28, 2024

Backdoor Attack

Papers

Imperceptible Rhythm Backdoor Attacks: Exploring Rhythm Transformation for Embedding Undetectable Vulnerabilities on Speech Recognition

Trading Devil: Robust backdoor attack via Stochastic investment models and Bayesian approach

Graph Neural Backdoor: Fundamentals, Methodologies, Applications, and Future Directions

Robustness-Inspired Defense Against Backdoor Attacks on Graph Neural Networks

A Survey of Backdoor Attacks and Defenses on Large Language Models: Implications for Security Measures

An LLM-Assisted Easy-to-Trigger Backdoor Attack on Code Completion Models: Injecting Disguised Vulnerabilities against Strong Detection

Lurking in the shadows: Unveiling Stealthy Backdoor Attacks against Personalized Federated Learning

Chain-of-Scrutiny: Detecting Backdoor Attacks for Large Language Models

PSBD: Prediction Shift Uncertainty Unlocks Backdoor Detection

Certified Robustness to Data Poisoning in Gradient-Based Training

Injecting Undetectable Backdoors in Obfuscated Neural Networks and Language Models

Mutual Information Guided Backdoor Mitigation for Pre-trained Encoders

BadAgent: Inserting and Activating Backdoor Attacks in LLM Agents

Unelicitable Backdoors in Language Models via Cryptographic Transformer Circuits

Invisible Backdoor Attacks on Diffusion Models

Generalization Bound and New Algorithm for Clean-Label Backdoor Attack

GANcrop: A Contrastive Defense Against Backdoor Attacks in Federated Learning

DiffPhysBA: Diffusion-based Physical Backdoor Attack against Person Re-Identification in Real-World

Cross-Context Backdoor Attacks against Graph Prompt Learning

Towards Unified Robustness Against Both Backdoor and Adversarial Attacks