avatar

Chenyang (Danny) Ma

PhD Candidate
University of Oxford
chenyang.ma [at] cs.ox.ac.uk


About Me

I am a third-year Ph.D. candidate in Computer Science at the University of Oxford, co-supervised by Prof. Andrew Markham and Prof. Niki Trigoni. My study is generously funded by the EPSRC DTP International Doctoral Scholarship. I had the privilege of collaborating closely with Xavier Puig and Ruta Desai at FAIR, Meta, and with Diego Romeres and Chiori Hori at MERL. I also spent wonderful times interning at IBM (2025), Roku (2024), and Flower Labs (2023).

Prior to my Ph.D., I worked as a research assistant in Cambridge Machine Learning Systems Lab, supervised by Prof. Nicholas Lane. I obtained my B.S.E. degree from University of Michigan, completing multiple projects under the guidances of Prof. Andrew Owens, Prof. Wenzhen Yuan, and Prof. Paul Green.

I enjoy collaborating on research and projects. Feel free to reach out if you are interested in a potential collaboration!

Research Interests

My research focuses on developing embodied (robotic) and AI agents in open-world settings. My Ph.D. work builds on three closely-related questions:

  1. How can we enhance agents’ reasoning and planning capabilities in dynamic environments, bridging the gap between visual perception, task descriptions, and low-level actions? I explore both improving the capabilities of VLMs/VLAs themselves and designing hierarchical agentic systems.
  2. How can agents self-correct or refactor their plans and actions, with or without human feedback?
  3. How can we enable agents that can collaborate with humans, accommodating individual preferences, safety needs, and spontaneous intentions?

On the side, I also enjoy working with vision applications that utilize various modalities (Vision, Tactile, 3D) and AI methods that are less centralized and more collaborative (Federated Learning).

News

Selected Publications (Full list on Google Scholar)

  1. ArXiv
    FurnitureVLA: Learning Long-Horizon Bimanual Furniture Assembly with Vision-Language-Action Model
    Chenyang Ma, Yue Yang, Radu Corcodel, Siddarth Jain, Andrew Wu, Chiori Hori, Diego Romeres
    ArXiv, 2026
    Project Page Paper

    ArXiv
    CycleVLA: Proactive Self-Correcting Vision-Language-Action Models via Subtask Backtracking and Minimum Bayes Risk Decoding
    Chenyang Ma, Kai Lu, Guangyu Yang, Jiuming Liu, Shitong Xu, Bill Byrne, Ioannis Havoutis, Niki Trigoni, Andrew Markham
    ArXiv, 2026
    Project Page Paper Code

    ArXiv
    CAIRN: Cross-Room 3D Scene Understanding with Topology-Aware Large Multimodal Models
    He Liang, Chenyang Ma, Yiming Zhang, Sangyun Shin, Andrew Markham, Niki Trigoni, Yuhang He
    ArXiv, 2026
    Project Page Paper Code

    ArXiv
    EmbeWebAgent: Embedding Web Agents into Any Customized UI
    Chenyang Ma, Clyde Fare, Matthew Wilson, Dave Braines
    Technical Report, 2026
    Project Page Paper Demo

    NeurIPS
    COOPERA: Continual Open-Ended Human-Robot Assistance
    Chenyang Ma, Kai Lu, Ruta Desai*, Xavier Puig*, Andrew Markham*, Niki Trigoni (*=Equal Advising)
    Neural Information Processing Systems (NeurIPS), 2025 Spotlight
    Project Page Paper Code

    ICRA
    KitchenVLA: Iterative Vision-Language Corrections for Robotic Execution of Human Tasks
    Kai Lu, Chenyang Ma, Chiori Hori, Diego Romeres
    Workshop on Safe-VLM (ICRA), 2025 Spotlight
    Paper

    NeurIPS
    SpatialPIN: Enhancing Spatial Reasoning Capabilities of Vision-Language Models through Prompting and Interacting 3D Priors
    Chenyang Ma, Kai Lu, Ta-Ying Cheng, Niki Trigoni, Andrew Markham
    Neural Information Processing Systems (NeurIPS), 2024
    Project Page Paper Code

    EuroMLSys
    Gradient-less Federated Gradient Boosting Tree with Learnable Learning Rates
    Chenyang Ma, Xinchi Qiu, Daniel Beutel, Nicholas Lane
    Workshop on Machine Learning and Systems (EuroMLSys), 2023
    Project Page Paper Code

    NeurIPS
    Touch and Go: Learning from Human-Collected Vision and Touch
    Fengyu Yang*, Chenyang Ma*, Jiacheng Zhang, Jing Zhu, Wenzhen Yuan, Andrew Owens (*=Equal Contribution)
    NeurIPS Datasets and Benchmarks Track, 2022
    Project Page Paper Dataset Code

    CVPR
    Sparse and Complete Latent Organization for Geospatial Semantic Segmentation
    Fengyu Yang*, Chenyang Ma* (*=Equal Contribution)
    Computer Vision and Patter Recognition (CVPR), 2022
    Paper

Industry Experience

  1. Mitsubishi
    Research Scientist Intern — Nov 2025 to Apr 2026
    Research Consultant — Apr 2025 to Sept 2025
    Advised by Diego Romeres, Chiori Hori

    Research Scientist Intern
    IBM — June 2025 to Sept 2025
    Advised by Dave Braines

    Research Collaborator
    FAIR, Meta — June 2024 to Apr 2025
    Advised by Xavier Puig, Ruta Desai

    Applied Scientist Intern
    Roku — July 2024 to Oct 2024
    Advised by Michael Sanders

    Research Scientist Intern
    Flower Labs — Oct 2022 to July 2023
    Advised by Nicholas Lane, Daniel Beutel

Services

Conference Reviewer

Teaching Assistant