Haichuan Zhang

Haichuan Zhang

PhD student in Computer Science at the University of Utah. Research interests: trustworthy machine learning, robustness, and generative models.

Haichuan Zhang

About Me

I am Haichuan Zhang, a Ph.D. student in the School of Computing at the University of Utah. My research focuses on trustworthy machine learning, particularly the security, robustness, and privacy of generative models and AI-driven systems. I am broadly interested in developing reliable, explainable, and safe ML systems that perform robustly in real-world settings.

Before joining the University of Utah, I obtained my Master’s and Bachelor’s degrees in Computer Science at Anhui University, where I developed a strong interest in generative models and AI security.

My current work investigates:

  • Backdoor attack and alignment on large language models
  • Robustness and defense of diffusion models
  • Broader security and safety issues in generative ai, and embodied ai systems

I enjoy building practical and theoretically grounded methods that enable trustworthy AI.

Selected Publications

  • Attack as Defense: Run-time Backdoor Implantation for Image Content Protection, arxiv preprint.
  • Are Your LLM-based Text-to-SQL Models Secure? Exploring SQL Injection via Backdoor Attacks, SIGMOD 2005.

Internships

iFlytek Research — Research Intern

Sep 2023 – Feb 2024

  • Designed datasets for LLM training
  • Deployed and optimized vision-language models
  • Result: One Chinese patent

Teaching

  • Practical Machine Learning, Teaching Assistant - Spring 2026
  • Database System Experimental Course, Teaching Assistant — Spring 2023

Service

  • Subreviewer for VLDB 2024, TIFS 2025, NeurIPS 2025, ICLR 2026

Skills

Languages: Python, C++, MATLAB, C
Frameworks: PyTorch, TensorFlow, Docker

Contact

hc.zhang@utah.edu