In recent years, Artificial Intelligence (AI) technology has developed rapidly and played an increasingly important role in the field of materials science. The data-driven AI4Science methods have become the core driving force of material simulation design because of its excellent balance between the inference speed and accuracy. In recent years, FUNSOM has innovated the research paradigm in the field of material design with AI technologies, and developed new AI-based methods for multi-scale atomic simulations and atomic structure, such as deep graph neural network and machine learning force field, which greatly improved the efficiency of material research and development. The structure, dynamic features, and various functional properties of molecules, polymers, nanomaterials and mesoscopic systems are studied and designed rationally from different scales.

In the direction of artificial intelligence and materials design, FUNSOM mainly focuses on the development of intelligent material simulation methods, intelligent material structure optimization, and intelligent material design. By conducting multi-scale simulations and research on functional nanomaterials and complex molecular systems, observing the phenomena and mechanisms in the simulations, rational and intelligent design of materials can be achieved.

1. Development of intelligent material simulation methods

Based on AI technology, such as machine learning force field, develop new theoretical simulation methods and multi-scale simulation methods to realize the simulation of complex processes of functional material systems, so as to explain experimental phenomena and mechanisms at the nano level.

2. Intelligent material structure optimization

Combining density functional theory, AI methods, such as reinforcement learning and Bayesian optimization, mesoscopic scale simulation methods and other methods to carry out automated research on functional nanomaterials and complex molecular systems, the microstructure and functional properties of materials can be efficiently optimized.

3. Intelligent material design

Build databases of functional nanomaterials, energy materials, etc., and give the relationship between material/molecular structures and their performance through AI method, such as graph neural network, so as to conduct in-depth study and intelligent screening of optoelectronic materials, energy materials, catalytic materials, biological materials, etc., and develop new materials.

Editor: Danting Xiang