A descriptor screening method, device and medium for multifunctional material development

By acquiring a set of material data, creating physical descriptors, and performing dimensionality reduction optimization and SHAP interpretation, important descriptors are selected, solving the problems of long development time and high cost in the development of multifunctional materials, and realizing efficient development of multifunctional materials and explicit relationship revelation.

CN119181450BActive Publication Date: 2026-06-26CHINA UNIV OF MINING & TECH
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
CHINA UNIV OF MINING & TECH
Filing Date
2024-09-18
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

In existing materials design research, the development of multifunctional materials is time-consuming and costly, and the optimization and iteration process of machine learning models requires repeated material performance tests, making it difficult to effectively reveal the explicit relationship between material structure and performance.

Method used

By acquiring a set of material data, physical descriptors are created, and dimensionality reduction optimization and SHAP algorithm interpretation are performed to select important descriptors as key physical factors for decoupling multiple properties, which are then used to evaluate the multifunctional properties of materials.

Benefits of technology

This enabled the screening of multifunctional materials at low cost and in a short time, revealed the explicit relationship between material structure and performance, improved design efficiency and reduced design costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119181450B_ABST
    Figure CN119181450B_ABST
Patent Text Reader

Abstract

The application discloses a descriptor screening method and device for multifunctional material development and a medium, relates to the field of material science, and comprises the following steps: creating a plurality of physical descriptors for the structure component information of a target material in a material data set and performing dimension reduction optimization; then inputting the optimized descriptor set into a plurality of target performance prediction models, and using a SHAP algorithm to interpret the models to obtain a plurality of optimized descriptor sorting coefficient sets; the target performance prediction model is a pre-trained machine learning model; based on the plurality of optimized descriptor sorting coefficient sets, the optimized descriptor set is subjected to feature sensitivity difference analysis and screening to obtain a plurality of important descriptors, which are key physical factors for decoupling between multiple performances of the target material. The application can comprehensively evaluate the numerical levels of different key physical factors of the material to guide the efficient development of multifunctional materials that can simultaneously meet multiple target performance indicators.
Need to check novelty before this filing date? Find Prior Art