Jian Lou (娄坚)

PreferCare: Preference Dataset Copyright Protection in LLM Alignment

PoisonedEye: Knowledge Poisoning Attack on RAG-based LVLMs

Assessing Safety Risks and Quantization-aware Safety Patching for Quantized LLMs

Cross-silo Federated Learning with Record-level Personalized DP

PromptCARE: Prompt Copyright Protection by Watermark Injection

DP-starJ + Local DP Heavy Hitter Detection in Data Streams

FINER: Enhancing Classifiers with Feature Attribution for Risk Analysis

Certified Minimax Unlearning with Generalization Rates

Equitable Data Valuation Meets the Right to be Forgotten

Dealer: An End-to-End Model Marketplace with Differential Privacy