Yicheng Duan

MSCS @ Case Western Reserve University. One click CV here.

prof_pic.jpg

Cleveland, 44120

Hi there! I’m Yicheng Duan (段毅成), currently pursuing my MS in Computer Science @ Case Western Reserve University, advised by Dr. Yu Yin. BS in Applied Physics @ University of Washington, Seattle.

My research interests center on building intelligent systems that can see, communicate, and act — especially humanoid agents capable of navigating, reasoning, and planning in complex environments, i.e. “Humanoid + world knowledge reasoning”. My research focuses on Vision-Language-Action models and Embodied AI. Lately my work has centered on efficient VLA, action representation, and benchmarking. I also bring hands-on experience with machine learning, vision-language navigation, VLM, and LLM.

Previously, as an algorithm engineer at ZHIPU.AI, I designed statistical and machine learning models, optimized large-scale data systems, contributed to the development of early-stage LLM agents, and enhanced backend performance. Inventor of 7+ patents.

The journey is still unfolding, and I’m always eager to connect, collaborate, and learn along the way.

“Making AI embodied, grounded, and real — together.”

news

Oct 20, 2025 Our work NEBULA — a unified ecosystem for vision–language–action agent evaluation — is now open-sourced. Visit the project page—we’d love your star: NEBULA.
Sep 24, 2025 A paper has been submitted to ICLR 2026.
Jul 15, 2025 Everything starts here. Stay tuned for the latest updates and announcements. 🦾

selected publications

  1. NEBULA: Do We Evaluate Vision-Language-Action Agents Correctly?
    Jierui Peng, Yanyan Zhang, Yicheng Duan, and 3 more authors
    2025
  2. Enhancing Subsequent Video Retrieval via Vision-Language Models (VLMs)
    Yicheng Duan, Xi Huang, and Duo Chen
    2025
  3. A Navigation Framework Utilizing Vision-Language Models
    Yicheng Duan and Kaiyu tang
    2025