Google Scholar
GitHub
About
Biography
I am a Senior Researcher at WeChat Vision, Tencent, working on multimodal foundation models. My research interest covers the full training stack — from large-scale pretraining to post-training optimization for models that perceive, reason, and act across image, video, audio, and language.
I received my Ph.D. from The Chinese University of Hong Kong, Shenzhen, advised by Prof. Ruimao Zhang and Prof. Zhen Li. During my Ph.D. (2021–2025), I was fortunate to spend research stays at IDEA, BIGAI, SenseTime Research, and Tencent, where I worked on multi-modal learning for diverse vision tasks.
▸Omni-modal Foundation Models
▸MLLM Pretraining & Post-Training
▸Video/Audio Understanding
📣 Hiring interns!
We are actively seeking self-motivated interns to work on related research topics, including
image / video / omni pretraining & post-training.
If you're interested, feel free to reach out via email!
Updates
News
- One paper is accepted to NeurIPS 2025.
- We present WeThink for general-purpose vision-language reasoning.
- One paper is accepted to T-PAMI.
- One paper is accepted to CVPR 2025.
- One paper is accepted to ICRA 2025.
- One paper is accepted to NeurIPS 2024.
- Grounding DINO is selected as The Most Influential Paper in ECCV 2024.
- Three papers are accepted to ECCV 2024.
- One paper is accepted to CVPR 2024.
- We present X-Pose to detect any keypoints of any objects.
- Grounded SAM is accepted to ICCV 2023 Demo Track.
- One paper is accepted to ICCV 2023.
- Two papers are accepted to MIDL 2023, one rated as oral presentation.
- One paper is accepted to CVPR 2023.
- One paper is accepted to ICLR 2023.
- One paper is accepted to NeurIPS 2022.
Research [Full list on Google Scholar]
Selected Publications
# corresponding author
Community
Academic Services
Conference & Journal Reviewer
Serving as a reviewer for top-tier venues in computer vision and machine learning:
CVPR
ICLR
ICML
NeurIPS
ECCV
ICCV
T-PAMI
TNNLS
TMM
Challenge Organizer
Co-organizer of the
MICCAI AMOS Segmentation Challenge 2022,
a large-scale benchmark on multi-organ abdominal segmentation across CT and MRI modalities.
Recognition
Honors & Awards
- 🏆The First Prize Scholarship · 2018, 2019, 2020
- 🏆National Scholarship · 2018