Publications

Zhengbang Yang, Yisheng (Eason) Zhong, Junyuan Hong, Zhuangdi Zhu (2026). CALIBURN: Self-Calibrated LLM Unlearning Alignment. In EMNLP 2026 Main Conference.

Yisheng (Eason) Zhong, Zhengbang Yang, Zhuangdi Zhu (2026). DUET: Distilled LLM Unlearning from an Efficiently Contextualized Teacher. In ICLR 2026.

Yisheng (Eason) Zhong, Yizhu Wen, Junfeng Guo, Mehran Kafai, Heng Huang, Hanqing Guo, Zhuangdi Zhu (2025). Web Intellectual Property at Risk: Preventing Unauthorized Real-Time Retrieval by Large Language Models. In EMNLP 2025 Main Conference.

Yisheng (Eason) Zhong, Li-Ping Wang (2024). PROFL: A Privacy-Preserving Federated Learning Method with Stringent Defense Against Poisoning Attacks. In CSCWD.