Overview: Artificial intelligence for computational physics aims to utilize computational techniques related to artificial intelligence to address complex and cutting-edge problems in physics. Research directions include the ground-state properties of atomic nuclei, the QCD phase diagram, jet quenching, and the non-equilibrium evolution of hot and dense nuclear matter. These areas involve the use of AI to solve complex differential equations, inverse problems, variational problems, and generative problems, as well as the application of physical principles to guide the design and optimization of AI algorithms.

Members：Wei-Bing Deng, Li-Ping Chi, De-Fu Hou, Chen Ji, Wei-Yao Ke, Wei Li, Xiao-Feng Luo, Zhi-Ming Li, Long-Gang Pang, Guang-You Qin, Xin-Nian Wang, Yuan-Fang Wu, Ming-Mei Xu, Ben-Wei Zhang, Han-Zhong Zhang.

The Research Center for Artificial Intelligence and Computational Physics at Central China Normal University is at the forefront of artificial intelligence algorithms, integrating advanced computational physics methods with high-performance parallel computing technologies. Our mission is to tackle historically insurmountable physical challenges and uncover unknown physical laws from vast datasets of particle and nuclear collisions. Our team, deeply rooted in the fundamental principles of physics, is dedicated to constructing precise digital twin systems for high-fidelity simulation of complex physical processes in large scientific facilities. Additionally, we employ advanced artificial intelligence techniques to extract key physical quantities from massive datasets, revealing hidden physical phenomena.

Our team members, drawn from diverse disciplines, possess expertise and skills in physics, computer science, mathematics, and more. Several large-scale physical simulation programs developed by our team have gained widespread international recognition and are openly available. These programs have successfully described numerous key observables in high-energy nuclear collisions, including:

• The HIJING model: Accurately simulates the production mechanisms of jets and mini-jets in nuclear collisions.

• The LBT and Lido programs: Detail the dynamical processes of high-energy partons traversing hot and dense nuclear matter.

• The CLVisc program: Simulates the spatiotemporal evolution of non-equilibrium quark-gluon plasma.

• The CoLBT program: Comprehensively describes the complex interactions of high-energy parton transport and medium response.

• The QLBT and Langevin dynamics models: Deeply analyze the transport behavior of heavy-flavor quarks in hot and dense nuclear matter.

Our research directions have achieved a series of cutting-edge international results at the intersection of computational physics and artificial intelligence. For instance, we employ random field Bayesian analysis techniques to determine jet energy loss and construct convolutional neural networks, point cloud neural networks, and graph convolutional neural networks to study the equation of state of nuclear matter and the QCD phase diagram. Furthermore, our team members have contributed to the review article "Machine Learning in Nuclear Physics" in the journal "Reviews of Modern Physics," which has had a profound impact in the international academic community, marking our leading position and innovative capabilities in this field.