Ming ZHANG

Ming Zhang 章銘

Ph.D.

Huawei Hong Kong Research Center,
Building 2W, Hong Kong Science Park, N.T.,
Hong Kong.

mzhang367-c (at) my (dot) cityu (dot) edu (dot) hk

I am currently a senior algorithm engineer in the Celia Team (Huawei's AI virtual assistant), at Huawei Hong Kong Research Center, in Pak Shek Kok, Hong Kong Science Park. Previously, I worked as a senior machine learning engineer at Hong Kong Applied Science and Technology Research Institute (ASTRI), and before that, I was a postdoctoral researcher at CIMDA.

I earned my Ph.D. in Electrical Engineering (EE), from City University of Hong Kong (CityU) in 2021, supervised by Prof. Hong Yan. Earlier, I obtained my MSc in Electronic Information Engineering (with Distinction), also from the EE department, CityU. I completed my B.Eng. in Automation from Nanjing Normal University in 2017.

I am now working on Multimodal LLM projects. Specifically, I am interested in unified multimodal models for image understanding, generation and editing. I am also interested in Continual Learning for MLLMs, aiming to preserve existing knowledge while acquiring new multimodal capabilities to achieve both stability and plasticity.

Selected Publications [Google Scholar]

(*) denotes corresponding author

  1. arXiv 2026
    Boogu-Image
    Boogu-Image-0.1: Boosting Open Agentic Multimodal Generation via Understanding under a Minimal Budget
    Core contributor, leading the few-step distillation team
    arXiv preprint, 2026
    paper project code model
  2. TPAMI 2026
    CUR Mix-Order Hypergraph Matching
    A CUR Decomposition-Based Mix-Order Framework for Large-Scale Hypergraph Matching
    Qixuan Zheng, Ming Zhang, Hong Yan.
    IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026
    paper
  3. CVPR 2024
    CURSOR
    CURSOR: Scalable Mixed-Order Hypergraph Matching with CUR Decomposition
    Qixuan Zheng, Ming Zhang*, Hong Yan.
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
    paper supp.
  4. PR 2023
    Orthonormal Product Quantization Network
    Orthonormal Product Quantization Network for Scalable Face Image Retrieval
    Ming Zhang*, Xuefei Zhe, Hong Yan.
    Pattern Recognition, 2023
    paper code
  5. PR 2021
    Deep Center-Based Dual-Constrained Hashing
    Deep Center-Based Dual-Constrained Hashing for Discriminative Face Image Retrieval
    Ming Zhang, Xuefei Zhe, Shifeng Chen, Hong Yan.
    Pattern Recognition, 2021
    paper code
  6. ICPR 2021
    Improved Deep Classwise Hashing
    Improved Deep Classwise Hashing With Centers Similarity Learning for Image Retrieval
    Ming Zhang, Hong Yan.
    International Conference on Pattern Recognition (ICPR), 2021
    paper
  7. PR 2020
    Deep Eigen-Filters
    Deep Eigen-Filters for Face Recognition: Feature Representation via Unsupervised Multi-Structure Filter Learning
    Ming Zhang, Sheheryar Khan, Hong Yan.
    Pattern Recognition, 2020
    paper