SaFo Lab @ JHU Johns Hopkins University

About

Chaowei Xiao

Current Position

Assistant Professor
Johns Hopkins University
Department of Computer Science

Faculty Researcher
NVIDIA

Email: chaoweixiao@jhu.edu

Biography

Chaowei Xiao is an Assistant Professor at Johns Hopkins University and a Faculty Researcher at NVIDIA. His research focuses on building safe AGI with both practical robustness and provable guarantees, as well as real-world applications including autonomous driving, agents, IoT, and healthcare. His recent research interests include safe and secure computer-use-agents and embodied agents.

Prior to joining Johns Hopkins, he was an Assistant Professor at the University of Wisconsin-Madison for two years. He received his Ph.D. from the University of Michigan, Ann Arbor, and his Bachelor’s degree from Tsinghua University.

Research Interests

  • LLM Security & Safety: Adversarial robustness, jailbreak attacks and defenses, prompt injection
  • Multimodal Learning: Vision-language models, multimodal LLMs
  • AI Safety: Trustworthy AI systems, certified robustness
  • Diffusion Models: Security and robustness of generative models
  • Autonomous Systems: Self-driving vehicles, embodied AI agents
  • Applied AI: IoT security, healthcare applications

Honors & Awards

  • 2024: Schmidt Sciences AI2050 Early Career Fellow
  • 2024: USENIX Security Distinguished Paper Award
  • 2024: Stanford/Elsevier Top 2% Scientists
  • 2025: ICLR Spotlight (AutoDAN-Turbo)
  • 2021: EWSN Best Paper Award
  • 2014: MobiCom Best Paper Award

Education

  • Ph.D. in Computer Science, University of Michigan, Ann Arbor
  • B.S. in Computer Science, Tsinghua University

Professional Service

  • Area Chair / Senior Program Committee: NeurIPS, ICLR, ICML, CVPR, ICCV, ECCV
  • Program Committee: ACL, EMNLP, NAACL, USENIX Security, IEEE S&P
  • Journal Reviewer: TPAMI, IJCV, JMLR

Selected Publications

Please visit our Publications page for a complete list of publications.

Lab Information

The Safe And secure Foundation mOdel systems Lab (SaFoLab) at Johns Hopkins University is dedicated to pioneering research in trustworthy (MultiModal) Large Language Model Systems. Our mission is to develop robust and secure AI systems that can be trusted across various application domains.

Lab Website: https://safo-lab.github.io
GitHub: https://github.com/SaFo-Lab

Last updated: January 2026