Mitigating Backdoor

Backdoor attacks, which surreptitiously embed malicious triggers into machine learning models during training, pose a significant threat to the reliability and security of AI systems. Current research focuses on developing robust defense mechanisms against these attacks, targeting various model architectures including large language models, federated learning systems, diffusion models, and vision transformers. These defenses employ diverse strategies such as weight manipulation, model editing, and trigger inversion, aiming to identify and neutralize backdoors without compromising the model's performance on legitimate data. The successful mitigation of backdoor attacks is crucial for ensuring the trustworthiness and widespread adoption of AI in sensitive applications.

Papers

November 17, 2024

BackdoorMBTI: A Backdoor Learning Multimodal Benchmark Tool Kit for Backdoor Defense Evaluation
Haiyang Yu, Tian Xie, Jiaping Gui, Pengyang Wang, Ping Yi, Yue Wu
Backdoor Attack Backdoor Defense Backdoor Learning Mitigating Backdoor Backdoor Model Multimodal Benchmark Datasets

October 18, 2024

DMGNN: Detecting and Mitigating Backdoor Attacks in Graph Neural Networks
Hao Sui, Bing Chen, Jiale Zhang, Chengcheng Zhu, Di Wu, Qinghua Lu, Guodong Long
Graph Neural Network Adversarial Attack Network Programming Graph Neural Stealthy Backdoor Attack Mitigating Backdoor Graph Backdoor Attack

September 30, 2024

Mitigating Backdoor Threats to Large Language Models: Advancement and Challenges
Qin Liu, Wenjie Mo, Terry Tong, Jiashu Xu, Fei Wang, Chaowei Xiao, Muhao Chen
Technical Challenge Backdoor Attack Hidden Backdoor Mitigating Backdoor

August 16, 2024

Mitigating Backdoor Attacks in Federated Learning via Flipping Weight Updates of Low-Activation Input Neurons
Binbin Ding, Penghui Yang, Zeqing Ge, Shengjun Huang
Backdoor Attack Weight Update Dead Neuron Mitigating Backdoor

July 31, 2024

Diff-Cleanse: Identifying and Mitigating Backdoor Attacks in Diffusion Models
Jiang Hao, Xiao Jin, Hu Xiaoguang, Chen Tianyou
Diffusion Model Backdoor Attack Backdoor Defense Backdoor Detection Mitigating Backdoor Backdoor Effect

July 10, 2024

Mitigating Backdoor Attacks using Activation-Guided Model Editing
Felix Hsieh, Huy H. Nguyen, AprilPyone MaungMaung, Dmitrii Usynin, Isao Echizen
Backdoor Attack Domain Generalization Backdoor Defense Mitigating Backdoor

May 25, 2024

Mitigating Backdoor Attack by Injecting Proactive Defensive Backdoor
Shaokui Wei, Hongyuan Zha, Baoyuan Wu
Backdoor Defense Backdoor Poisoning Attack Mitigating Backdoor

March 6, 2024

On the Effectiveness of Distillation in Mitigating Backdoors in Pre-trained Encoder
Tingxu Han, Shenghan Huang, Ziqi Ding, Weisong Sun, Yebo Feng, Chunrong Fang, Jun Li, Hanwei Qian, Cong Wu, Quanjun Zhang, Yang Liu, Zhenyu Chen
Pre Trained Mutual Distillation Low Temperature Distillation Distillation Loss Mitigating Backdoor

November 13, 2023

Mitigating Backdoors within Deep Neural Networks in Data-limited Configuration
Soroush Hashemifar, Saeed Parsa, Morteza Zakeri-Nasrabadi
Deep Neural Network DNN Framework Limited Data Mitigating Backdoor Backdoor Policy

July 28, 2023

Beating Backdoor Attack at Its Own Game
Min Liu, Alberto Sangiovanni-Vincentelli, Xiangyu Yue
Backdoor Attack Adversarial Backdoor Mitigating Backdoor

July 2, 2023

FedDefender: Backdoor Attack Defense in Federated Learning
Waris Gill, Ali Anwar, Muhammad Ali Gulzar
Malicious Client DEfense Mechanism Mitigating Backdoor Differential Testing

June 3, 2023

Mitigating Backdoor Attack Via Prerequisite Transformation
Han Gao
Backdoor Attack Mitigating Backdoor Backdoor Model

March 1, 2023

Mitigating Backdoors in Federated Learning with FLD
Yihang Lin, Pengyuan Zhou, Zhiqian Wu, Yong Liao
Federated Learning Backdoor Attack Mitigating Backdoor Backdoor Model

October 18, 2022

Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models
Zhiyuan Zhang, Lingjuan Lyu, Xingjun Ma, Chenguang Wang, Xu Sun
Natural Language Processing Backdoor Attack Backdoor Defense Fine Tuned Language Model Mitigating Backdoor

June 24, 2022

Defending Backdoor Attacks on Vision Transformer via Patch Processing
Khoa D. Doan, Yingjie Lao, Peng Yang, Ping Li
Convolutional Neural Network Vision Transformer Adversarial Example Backdoor Attack Backdoor Injection Mitigating Backdoor

May 26, 2022

Circumventing Backdoor Defenses That Are Based on Latent Separability
Xiangyu Qi, Tinghao Xie, Yiming Li, Saeed Mahloujifar, Prateek Mittal
Backdoor Defense Backdoor Poisoning Attack Backdoor Learning Mitigating Backdoor Latent Separation