Shengcao Cao.

Research Engineer 曹胜操

Shengcao
Cao

I build visual intelligence that learns with less human supervision.

Google DeepMind · Ph.D., University of Illinois Urbana-Champaign

Shengcao Cao with his two cats, Kaka and Dola Figure 1. Left: Kaka “好运来”, a five-year-old orange tabby. Right: Dola “多乐”, a one-year-old ragdoll. Middle: their human, Shengcao♥

About

Learning visual intelligence with less human supervision.

I am a Research Engineer at Google DeepMind and a Ph.D. candidate in Computer Science at the University of Illinois Urbana-Champaign, advised by Liangyan Gui and Yuxiong Wang, graduating in August 2026. Earlier, I received my M.S. in Robotics from Carnegie Mellon University, working with Kris Kitani, and my B.S. in Computer Science from Peking University, working with Liwei Wang.

My research builds visual intelligence that learns with less human supervision, spanning self-supervised learning, open-world detection and segmentation, and large multimodal models. I work toward autonomous models and agents that discover knowledge and structure on their own, and toward omni-modal representations that integrate vision, language, and other modalities into a shared understanding of the world.

Experience

  • May 2026 — Present

    Research Engineer

    Google DeepMind

    Mountain View, CA

  • May 2025 — Dec 2025

    Student Researcher

    Google DeepMind

    with Tanmaya Dabral & Zhongli Ding

  • May 2024 — Nov 2024

    Applied Research Scientist Intern

    Adobe Research

    with Zijun Wei & Jason Kuen

  • May 2023 — Aug 2023

    Research Scientist Intern

    Adobe Research

    with Jiuxiang Gu & Jason Kuen

  • May 2022 — Aug 2022

    Visiting Scholar

    IBM Research — T. J. Watson Center

    with Dhiraj Joshi & Nirmit Desai

Education

  • Aug 2021 — Aug 2026

    Ph.D., Computer Science

    University of Illinois Urbana-Champaign

    advised by Liangyan Gui & Yuxiong Wang

  • Aug 2019 — May 2021

    M.S., Robotics

    Carnegie Mellon University

    advised by Kris Kitani

  • Sep 2015 — Jul 2019

    B.S., Computer Science

    Peking University

    advised by Liwei Wang

Selected publications

  1. Think in Latent, Explain in Language: Self-Explainable Latent Reasoning

    Dayuan Zhao*, Shengcao Cao*, Yu-Xiong Wang, Liang-Yan Gui

    ICML 2026

  2. CoCo-IR: Contextual Composed Image Retrieval

    Shengcao Cao, Tanmaya Shekhar Dabral, Zhongli Ding, Madhuri Shanbhogue, Kaifeng Chen, Zhe Li, Mojtaba Seyedhosseini, Liang-Yan Gui, Yu-Xiong Wang

    ECCV 2026

  3. Refer to Anything with Vision-Language Prompts

    Shengcao Cao, Zijun Wei, Jason Kuen, Kangning Liu, Lingzhi Zhang, Jiuxiang Gu, HyunJoon Jung, Liang-Yan Gui, Yu-Xiong Wang

    ICCV 2025

  4. Emergent Visual Grounding in Large Multimodal Models Without Grounding Supervision

    Shengcao Cao, Liang-Yan Gui, Yu-Xiong Wang

    ICCV Findings 2025

  5. Aligning Large Multimodal Models with Factually Augmented RLHF

    Zhiqing Sun*, Sheng Shen*, Shengcao Cao*, Haotian Liu, Chunyuan Li, Yikang Shen, Chuang Gan, Liang-Yan Gui, Yu-Xiong Wang, Yiming Yang, Kurt Keutzer, Trevor Darrell

    ACL Findings 2024

  6. SOHES: Self-Supervised Open-World Hierarchical Entity Segmentation

    Shengcao Cao, Jiuxiang Gu, Jason Kuen, Hao Tan, Ruiyi Zhang, Handong Zhao, Ani Nenkova, Liang-Yan Gui, Tong Sun, Yu-Xiong Wang

    ICLR 2024

  7. HASSOD: Hierarchical Adaptive Self-Supervised Object Detection

    Shengcao Cao, Dhiraj Joshi, Liang-Yan Gui, Yu-Xiong Wang

    NeurIPS 2023

  8. Learning Lightweight Object Detectors via Progressive Knowledge Distillation

    Shengcao Cao, Mengtian Li, James Hays, Deva Ramanan, Yu-Xiong Wang, Liang-Yan Gui

    ICML 2023

  9. Contrastive Mean Teacher for Domain Adaptive Object Detectors

    Shengcao Cao, Dhiraj Joshi, Liang-Yan Gui, Yu-Xiong Wang

    CVPR 2023

  10. Rethinking Transformer-Based Set Prediction for Object Detection

    Zhiqing Sun*, Shengcao Cao*, Yiming Yang, Kris M. Kitani

    ICCV 2021

* denotes equal contribution.  •  Full list on Google Scholar.