Mr. Zeren Zhang, Peking University, China
Zeren Zhang, a 27-year-old researcher from Sichuan, China, is currently pursuing a Ph.D. in Applied Mathematics at Peking University, specializing in multimodal large models, large language models, and digital human generation, with a GPA of 3.77/4. He earned his Bachelor’s degree in Statistics from Sichuan University, where he ranked first in his major with a GPA of 3.78/4 and received the prestigious National Scholarship. Zeren’s professional experience includes roles as an Algorithm Researcher at 01Wanwu and Baidu and an Algorithm Engineer at NetEase Youdao, contributing to groundbreaking projects like DFE-GPS, SwapTalk, and transfer learning for GANs. His research innovations have been published in prestigious journals and conferences such as ICASSP, ECMLPKDD, and Remote Sensing. Proficient in PyTorch and advanced AI methodologies, Zeren has made significant strides in developing cutting-edge technologies, including geometry problem solvers, audio-driven talking-face models, and seismic fault detection systems.
Professional Profile:
ORCID
Summary of Suitability for the Best Researcher Award:
Zeren Zhang, a 27-year-old Ph.D. candidate at Peking University specializing in Applied Mathematics, demonstrates exceptional research prowess in multimodal large models, large language models, and digital human generation. His academic and professional accomplishments make him a strong contender for the Best Researcher Award.
🎓 Education
- Peking University
- Degree: Ph.D. in Applied Mathematics
- Duration: September 2020 – Present
- Research Focus: Multimodal Large Models, Large Language Models, Digital Human Generation
- Achievement: GPA: 3.77/4
- Sichuan University
- Degree: Bachelor’s in Statistics
- Duration: September 2016 – June 2020
- Achievement:
- GPA: 3.78/4 (Ranked 1st in Major) 🥇
- Recipient of the National Scholarship (2019–2020 Academic Year) 🏆
💼 Work Experience
- 01Wanwu
- Role: Algorithm Researcher
- Duration: March 2024 – August 2024
- Achievement:
- Developed DFE-GPS, a multimodal geometry problem-solving model with an 82.38% accuracy on the FormalGeo dataset, surpassing GPT-4. 🖥️
- Published in ICASSP 2025. 📖
- NetEase Youdao
- Role: Algorithm Engineer
- Duration: October 2023 – March 2024
- Achievement:
- Designed and trained the SwapTalk model for audio-driven, customizable talking-face video generation. 📹
- Integrated into Youdao’s digital human business and published in ICASSP 2025. 📖
- Baidu
- Role: Algorithm Researcher
- Duration: September 2022 – May 2023
- Achievement:
- Researched transfer learning for pretrained GANs using domain-adaptive Gaussian Mixture Models (GMM).
- Published in ECMLPKDD 2023. 📖
🏅 Awards and Honors
- National Scholarship (2019–2020 Academic Year) 🏆
- Ranked 1st in Major during Bachelor’s studies at Sichuan University 🥇
Publication Top Notes:
Improving Seismic Fault Recognition with Self-Supervised Pre-Training: A Study of 3D Transformer-Based with Multi-Scale Decoding and Fusion
SG-Net: Semantic Guided Network for Image Dehazing