About
This is Hanyu Zhu. I am a Ph.D. student in Computer Science at University of Massachusetts Dartmouth. I obtained my bachelor's degree from Sichuan University.
My current research interests mainly lie in security and privacy and trustworthy AI systems. In particular, I focus on (1) LLM/MLLM Security: attack, defense, and safety alignment; (2) Trustworthy Cyber-Physical Systems (CPS): vulnerabilities analysis and robust design for large-model-enabled systems with physical world intersections, such as Integrated Sensing and Communications (ISAC).
I am also interested in interdisciplinary AI/Cybersecurity themes. I actively welcome collaboration opportunities. Feel free to reach out to me. ✉️
News
[09/2025] 🎉 One paper has been accepted by NeurIPS 2025.
Recent Publications
- Hanyu Zhu, Lance Fiondella, Jiawei Yuan, Kai Zeng, and Long Jiao. "NeuroGenPoisoning: Neuron-Guided Attacks on Retrieval-Augmented Generation of LLM via Genetic Optimization of External Knowledge", the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025). [paper]