Me
Junyao Gao (高俊尧)
Phd Candidate @ Tongji

About

Hello, I am a third-year Computer Science Ph.D. student at Tongji University, luckily advised by Prof. Cairong Zhao. Prior to this, I obtained my bachelor degree in Intelligence Science and Technology from NJUST in 2022. At present, I am doing research intern in Tencent (AIPD), Shenzhen, supervised by Jun Zhang and Sibo Liu. Before that, I spent a happy journey in OpenMMLab of Shanghai AI Laboratory, Shanghai, supervised by Dr. Yanan Sun. I also have a deep academic collaboration with Dr. Xinyang Jiang from Microsoft Research Asia. My research interests are in image/video generation, privacy & security and person Re-ID, etc.

Education


Ph.D.           Sep. 2022 - Present
                    Tongji University, Shanghai, China.
                    Ph.D. student in Computer Science
B.S.             Sep. 2018 - Jun. 2022
                    NJUST , Nanjing, China.
                    B.S.(Honors) in Intelligence Science and Technology

Experiences

  • Jun. 2025 -- Present, Tencent, Shenzhen, China
         Research Intern, work on image/video generation.
  • Nov. 2023 -- May. 2025, Shanghai AI Lab, Shanghai, China
         Research Intern, work on image/video generation.

    • News

    • We release MegaStyle, which is a novel and scalable data curation pipeline that first explores consistent T2I style mapping ability from current large generative models to construct intra-style consistent, inter-style diverse and high-quality style dataset.
    • One papers (DiffPano++) about multi-view panorama generation are accepted to IJCV 2026.
    • Two papers (DynamicsBoost, DiT-Distill) about plausible video generation and fine-grained retrieval are accepted to CVPR 2026.
    • Two papers (AMD, TiMi) about dataset distillation and agentic system for quantitative financial trading are accepted to ICLR 2026.
    • StyleShot is accepted by IEEE TPAMI 2025!!!
    • We release CharacterShot, which supports diverse character designs and custom motion control (2D pose sequence), enabling 4D character animation in minutes and without specialized hardware.
    • One papers (CamPVG) about panorama generation accepted to SIGGRAPH ASIA 2025.
    • Two papers (MotionShot, CrossVLAD) about motion transfer in video generation and adversarial attack on unified vision-language models are accepted to ICCV 2025.
    • I will be joining the AIPD at Tencent (IP-Adapter Team) to begin my internship journey. It's great to see you in Shenzhen!
    • One paper (MCDM) about long-term talkingface generation accepted to ICML 2025.
    • We release LEGO-Puzzles, a new benchmark with 1,100 LEGO-based VQA tasks that evaluates MLLMs' ability to perform multi-step spatial reasoning.
    • One paper (UniDD) about dataset distillation accepted to CVPR 2025 as the oral presentation.
    • One paper (FaceShot) about portrait animation accepted to ICLR 2025.
    • One paper (DiffPano) about panorama generation accepted to NIPS 2025.
    • We release StyleShot, a style transfer method capable of generating the high-quality stylized images that match the desired style from any reference image without test-time style-tuning.
    • One paper (Re-ID leak) about MIA in Re-ID accepted to IJCV.
    • One paper about fine-grained classification accepted to AAAI 2024.
    • One paper (SDMIA) about MIA in Re-ID accepted to AAAI 2023 as the oral presentation.

    Selected Publications

    MegaStyle: Constructing Diverse and Scalable Style Dataset via Consistent Text-to-Image Style Mapping
    Junyao Gao*, Sibo Liu*, Jiaxing Li, Yanan Sun, Yuanpeng Tu, Fei Shen, Weidong Zhang, Cairong Zhao, Jun Zhang (* equal contribution)
    Preperint, 2026
    [ paper] [project page]
    CharacterShot: Controllable and Consistent 4D Character Animation.
    Junyao Gao*, Jiaxing Li*, Wenran Liu, Yanhong Zeng, Fei Shen, Kai Chen, Yanan Sun, Cairong Zhao (* equal contribution)
    Preperint, 2025
    [ paper] [code]

    LEGO-Puzzles: How Good Are MLLMs at Multi-Step Spatial Reasoning?
    Kexian Tang*, Junyao Gao*, Yanhong Zeng, Haodong Duan, Yanan Sun, Zhening Xing, Wenran Liu, Kaifeng Lyu, Kai Chen (* equal contribution)
    Preperint, 2025
    [ paper] [code] [project page]

    FaceShot: Bring Any Character into Life
    Junyao Gao, Yanan Sun, Fei Shen, Xin Jiang, Zhening Xing, Kai Chen, Cairong Zhao,
    ICLR 2025
    [ paper] [code] [project page]

    StyleShot: A Snapshot on Any Style
    Junyao Gao, Yanchen Liu, Yanan Sun, Yinhao Tang, Yanhong Zeng, Kai Chen, Cairong Zhao
    TPAMI, 2025
    [ paper] [code] [project page] [demo]

    Re-ID-leak: Membership Inference Attacks Against Person Re-identification
    Junyao Gao, Xinyang Jiang, Shuguang Dou, Dongsheng Li, Duoqian Miao, Cairong Zhao
    IJCV, 2024
    [ paper] [code]

    Similarity distribution based membership inference attack on person re-identification
    Junyao Gao, Xinyang Jiang, Huishuai Zhang, Yifan Yang, Shuguang Dou, Dongsheng Li, Duoqian Miao, Cheng Deng, Cairong Zhao
    AAAI (oral), 2023
    [ paper] [code]

    Selected Honors & Awards

    • Outstanding Undergraduate in Terms of Academic Performance, 2022
    • National Scholarship (Top 1%), 2021
    • Outstanding Student Cadres in Jiangsu Province, 2021

    Academic Services

    Conference Reviewer: NeurIPS, ICLR, ICML, ICCV, CVPR, ECCV, AAAI, ACM MM, WACV, AISTATS, BMVC
    Journal Reviewer: IJCV

    Misc

    I am willing to try something new, like guitar, photography and writing poem, and always full of novel ideas.
    Enjoy participating in meaningful social activities (connect with the community).
    I am also used to going to the gym in the evening, and running or swimming in the morning.