My name is Wei Dai (戴威), I’m a 4th-year undergraduate student at the School of Communication Engineering, Xidian University, supervised by Yanbiao Ma. Meanwhile, I’m a 4th-year undergraduate student at the School of Electronic Engineering, Heriot-Watt University, supervised by João Mota. Furthermore, I am a direct PhD student of the 2026 cohort at Tsinghua University,supervised by Shao-lun Huang.

🌟 Newest update:

(1) I am currently interning at Tsinghua TBSI, focusing on efficient transfer learning(4LLM). I look forward to sharing new research findings soon.

(2) I am completing my graduation project for Federated Graph Learning under the guidance of João Mota, who is a patient and nice teacher.

(3) Plan some interesting entrepreneurial projects with some passionate partners (4MultiAgent).

🔎 Research

“The laws of the physical world serve as the fundamental driving force behind the advancement of artificial intelligence”

I am passionate about using physical Laws to explain principle in the world. Currently, my research interests mainly focus on the following aspects:

a) Multimodal Large Model

b) Transfer Learning

c) Trustworthy Machine Learning


🔥 News

  • 2026.2: 🚀 One paper is accepted by TPAMI.
  • 2026.2: 🚀 One paper is accepted by ICLR.
  • 2025.12: ✨ I serve as a reviewer for ICLR 2026.
  • 2025.3: 🚀 One paper is accepted by AAAI. See you in Singapore.
  • 2025.3: 🌐 2025CVPR_GGEUR Code is publicly available! Office-Home-LDS Dataset is publicly available!
  • 2025.3: 🚀 One paper is accepted by CVPR(5/5/5/4) Oral. See you in Tennessee.
  • 2025.2: 🚀 One paper is accepted by ICLR. See you in Singapore.
  • 2024.12: ✨ I serve as a reviewer for CVPR 2025.
  • 2024.10: 🎈 I was honored with a scholarship from Xidian University.


📃 Publications

† Equal Contribution

Geometric Knowledge-Guided Localized Global Distribution Alignment for Federated Learning [Github] [Dataset]
Yanbiao Ma†, Wei Dai† (co-first), Wenke Huang, Jiayi Chen
Conference on Computer Vision and Pattern Recognition (CCF-A CVPR), Oral Presentation (Top 0.5%), 2025

Calibrating Biased Distribution in VFM-derived Latent Space via Cross-Domain Geometric Consistency [Github]
Yanbiao Ma, Wei Dai, Bowei Liu, Jiayi Chen, Wenke Huang, Guancheng Wan, Zhiwu Lu, Junchi Yan
IEEE Transactions on Pattern Analysis and Machine Intelligence (CCF-A TPAMI), 2026

Pursuing Better Decision Boundaries for Long-Tailed Object Detection via Category Information Amount
Yanbiao Ma, Wei Dai, Jiayi Chen
International Conference on Learning Representations (Tsinghua-A ICLR), 2025

Compositional Attribute Imbalance in Vision Datasets
Yanbiao Ma†, Jiayi Chen†, Wei Dai, Dong Zhao, Zeyu Zhang, Yuting Yang, Bowei Liu, Jiaxuan Zhao, Andi Zhang
AAAI Conference on Artificial Intelligence (CCF-A AAAI), 2025

🎖 Awards

  • ✨✨✨ Academician Duan Baoyan's Science and Education Innovation Award.
  •  
  • ✨✨✨ Xidian University scholarship.

🎡 Service

Conference Committee Member

  • Reviewer for ICLR’2026
  • Reviewer for ICML’2026
  • Reviewer for ICLR’2025
  • Reviewer for CVPR’2025

    Journal Reviewer

  • Reviewer for IEEE Access

📖 Educations

  • 2026.06 - Now, Phd, Tsinghua Shenzhen International Graduate School, Tsinghua University, China
  • 2022.09 - 2026.06, Undergraduate, School of Communication Engineering, Xidian University, China (GPA:3.9/4.0, Weighted Average scorc:91/100)
  • 2022.09 - 2026.06, Undergraduate, School of Engineering and Physical Sciences, Heriot-Watt University, UK