Hongyi Wen

(pronounced as: HongE When)

prof_pic.jpg

🏢S712, Qiantan Campus

I’m an Assistant Professor of Computer Science at NYU Shanghai. Prior to that, I received my Ph.D. in Information Science from Cornell University advised by Deborah Estrin. I hold a B.Eng in Computer Science from Tsinghua University.

My research interest lies in developing human-in-the-loop machine learning systems and improving long-term values of personalizations. I lead the Multimodal Agentic Personalization Systems (MAPS) research group at NYU Shanghai. Our recent work focuses on generative models and their applications in personalizations.

📢 I’m looking for PhDs to join our research group. Check our PhD program and contact me at hongyi.wen@nyu.edu if you are interested in the following topics:

  • Understanding generative models(LLM, Diffusion, etc) and making them more personalized for individuals
  • AI for Learning and Creativity

selected publications

  1. ACL
    SHAPE: Unifying Safety, Helpfulness and Pedagogy for Educational LLMs
    Zhao, Sihang, Yu, Kangrui, Yuan, Youliang, He, Pinjia, and Wen, Hongyi
    ACL Oral 2026
  2. CVPR
    From Navigation to Refinement: Revealing the Two-Stage Nature of Flow-based Diffusion Models through Oracle Velocity
    Liu, Haoming, Liu, Jinnuo, Li, Yanhao, Bai, Liuyang, Ji, Yunkai, Guo, Yuanhe, Wan, Shenjie, and Wen, Hongyi
    CVPR Findings 2026
  3. ICCV
    ImageGem: In-the-wild Generative Image Interaction Dataset for Generative Model Personalization
    Guo, Yuanhe, Xie, Linxi, Chen, Zhuoran, Yu, Kangrui, Po, Ryan, Yang, Guandao, Wetztein, Gordon, and Wen, Hongyi
    2025
  4. EMNLP
    Reveal and Release: Iterative LLM Unlearning with Self-generated Data
    Xie, Linxi, Teng, Xin, Ke, Shichang,  Wen, Hongyi, and Wang, Shengjie
    2025
  5. WSDM
    GEMRec: Towards Generative Model Recommendation
    Guo, Yuanhe, Liu, Haoming, and Wen, Hongyi
    WSDM 2024