Chunlin Ji
Kuang-Chi Institute of Advanced Technology, Shenzhen, 518000, China 

EXTENDED ABSTRACT: The scientific problem of synergistic optimization of multiple properties of functional materials and decision-making on material composition has been investigated. Research on multi-objective optimization algorithms has been conducted, particularly proposing active learning methods that integrate multi-objective optimization for small-sample material data. Uncertainty quantification methods applicable to functional material data modeling have been explored, and a multi-objective active learning method that incorporates uncertainty quantification has been proposed. A material design software framework integrating data processing, machine learning, and optimization algorithms has been developed, forming a material performance multi-objective optimization decision-making system that includes the "data processing - model construction - optimization decision" process. Requirements analysis and overall architecture design of an intelligent software platform for multi-objective synergistic optimization of functional materials have been conducted, proposing software design methods such as cross-domain trusted services and visualization modeling optimization process design. An algorithm platfom飞ith machine learning modeling and optimization methods as the core and a data platform with data processing and characterization as the core have been designed. The intelligent software platform has been preliminarily implemented, and this report will provide application demonstrations of the software platform.
Keywords: Intelligent platform; Software platform; Functional material design; Multi objective collaborative optimization; Quantification of uncertainty; Machine learning; Data driven