Pan Li

4 papers Georgia Institute of Technology

• Differentially Private Relational Learning with Entity-level Privacy Guarantees
• Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness
• Graph-KV: Breaking Sequence via Injecting Structural Biases into Large Language Models
• RoFt-Mol: Benchmarking Robust Fine-tuning with Molecular Graph Foundation Models