-
Patcher: Post-Hoc Patching of Backdoored Large Language Models
PDFCodeAnjun Gao, Yueyang Quan, Yufei Xia, Zhuqing Liu, and Minghong Fang
In Proc. USENIX Security Symposium, 2026
-
Five Queries Are Enough: Query-Efficient and Surrogate-Free Membership Inference Attacks on RAG via Entailment
PDFCodeNguyen Linh Bao Nguyen, Wanlun Ma, Viet Vo, Alsharif Abuadbba, Minghong Fang, Jun Zhang, and Yang Xiang
In Proc. USENIX Security Symposium, 2026
-
Who Taught the Lie? Responsibility Attribution for Poisoned Knowledge in Retrieval-Augmented Generation
PDFPosterCodeBaolei Zhang*, Haoran Xin*, Yuxi Chen, Zhuqing Liu, Biao Yi, Tong Li, Lihai Nie, Zheli Liu, and Minghong Fang
In Proc. IEEE Symposium on Security and Privacy, 2026
(*co-primary authors, acceptance rate: 12.7%)
-
Do We Really Need to Design New Byzantine-robust Aggregation Rules?
PDFMinghong Fang, Seyedsina Nabavirazavi, Zhuqing Liu, Wei Sun, Sundararaja Sitharama Iyengar, and Haibo Yang
In Proc. NDSS, 2025
(acceptance rate: 16.1%)
-
Byzantine-Robust Decentralized Federated Learning
PDFMinghong Fang, Zifan Zhang, Hairi, Prashant Khanduri, Jia Liu, Songtao Lu, Yuchen Liu, and Neil Gong
In Proc. ACM CCS, 2024
(acceptance rate: 16.9%)
-
FLTrust: Byzantine-robust Federated Learning via Trust Bootstrapping
PDFCodeXiaoyu Cao*, Minghong Fang*, Jia Liu, and Neil Zhenqiang Gong
In Proc. NDSS, 2021
(*co-primary authors, acceptance rate: 15.2%)
-
Local Model Poisoning Attacks to Byzantine-Robust Federated Learning
PDFCodeMinghong Fang*, Xiaoyu Cao*, Jinyuan Jia, and Neil Zhenqiang Gong
In Proc. USENIX Security Symposium, 2020
(*co-primary authors, acceptance rate: 16.1%)