​Dr. Zhengde Zhang, a distinguished talent in high-energy physics, serves as an associate researcher and master's supervisor. His primary research focuses on AI algorithms, large models, and intelligent agents aimed at scientific discovery. He holds a Ph.D. in Particle and Nuclear Physics and has completed postdoctoral research in deep learning algorithms. In 2022, he was recruited by the Institute of High Energy Physics, where he leads projects such as the "From 0 to 1 Project - Research on AI Large Models Driving Scientific Discovery in High-Energy Physics" and "Research and Demonstration of High-Energy Physics Big Data Technology Based on Artificial Intelligence." He has developed neural networks such as CDNet, FINet, and MWNet, and has been instrumental in planning and building the High-Energy Physics Artificial Intelligence platform, HepAI, accessible at https://ai.ihep.ac.cn. Furthermore, he has advanced the HEP·Xiwu large language model and the "Dr. Sai" scientific research intelligent agents, dedicated to leveraging AI technology to facilitate discoveries in high-energy physics.

Deep learning algorithms;

Large model for scientific data;

Artificial intelligence platforms and software system;

Next generation AI algorithms driven by knowledge and data

09/2015 - 06/2020 Ph.D. in Particle Physics and Nuclear Physics, Shanghai Institute of Applied Physics, Chinese Academy of Sciences.

09/2011 - 06/2015 B.S. in Applied Physics, School of Physics, Beihang University.

07/2022 - Now       Associate researcher, talent introduction, Institute of High Energy Physics, Chinese Academy of Sciences.

07/2020 - 06/2022 Postdoctoral in artificial intelligence engineering at the Information and Communication Engineering Mobile Station of Shanghai Jiao Tong University.

(1)     AIDNet: Detecting Insulators and Defects from Satellite Remote Sensing Images: An Exploration, 2022 International Conference on Computer Networks and Communications, 2023, Corresponding author

(2)     MWNet: A Tracking Method for Frequently Occluded Scenes Based on Matter Waves, 2022 IEEE International Conference on Image Processing (ICIP), 2022, Third author

(3)     CDNet: a Real-time and Robust Crosswalk Detection Network on Jetson Nano Based on YOLOv5, NEURAL COMPUTING & APPLICATIONS, 2022, First Author

(4)     Burnup Optimization of Once-through Molten Salt Reactors using Enriched Uranium and Thorium, NUCLEAR SCIENCE AND TECHNIQUES, 2022, Third author

(5)     FINet: An Insulator Dataset and Detection Benchmark Based on Synthetic Fog and Improved YOLOv5, IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2022, First author

(6)     A Transmission Line and Its Defect Detection Method Based on Data Enhancement, Augmentation and Neural Network, SOUTHERN POWER SYSTEM TECHNOLOGY, 2022, Corresponding author

(7)     High-throughput Screening and Machine Learning for the Efficient Growth of High-quality Single-wall Carbon Nanotubes, NANO RESEARCH, 2021, Sixth author

(8)     Migration Behavior of Tellurium in Bcc Iron Against Typical Alloying Elements: A First-principles Study, COMPUTATIONAL MATERIALS SCIENCE, 2020, First author

(9)     Direct Synthesis of Semimetal Phthalocyanines on a Surface with Insights into Interfacial Properties, JOURNAL OF PHYSICAL CHEMISTRY C, 2020, Fourth author

(10)  VaspCZ: an Efficient VASP Computation Assistant Program, NUCLEAR SCIENCE AND TECHNIQUES (Chinese), 2020, First author

(11)  First-principles Study of Helium Behavior in Nickel with Noble Gas Incorporation, JOURNAL OF APPLIED PHYSICS, 2020, Fourth author

(12)  Ullmann Coupling of 2,7-dibromopyrene on Au(111) Assisted by Surface Adatoms, APPLIED SURFACE SCIENCE, 2020, Third author

Zhengde Zhang developed the high-throughput VASP computing aid program VaspCZ to improve efficiency. He designed and implemented real-time robust neural network CDNet for complex weather conditions, FINet and TLDNet for small samples and extremely low-resolution conditions, respectively. He proposed a new AI algorithm driven by knowledge and data collaboration based on quantum wave functions and preliminarily validated in dense occlusion scenarios. He is developing ChatHEP, a fine-tuned large language model in the field of high-energy physics.

He has published more than ten papers in important academic journals both domestically and internationally, including Neural Compute Appl, IEEE T Instrument Meas, Nano Research, Comp Mate Sci, etc. He has opened up four projects on Github and racked up nearly 300 stars. Three related patents and one software copyright have been approved, and one technical achievement transfer has been completed.

Open sourced projects：github.com/zhangzhengde0225

​

Open source projects: github.com/zhangzhengde0225

AI platform: ai.ihep.ac.cn

现指导学生

唐梓涵  硕士研究生  085404-计算机技术  

张天硕  硕士研究生  085404-计算机技术