AsianScientist (Nov. 21, 2018) – Researchers at the National Institute of Materials Science (NIMS), Japan, and the Toyota Technological Institute at Chicago, US, have jointly developed a computer-aided material design (CAMaD) system capable of extracting information vital to material design. Their research is published in Science and Technology of Advanced Materials.
The performance of a material is determined by its properties, which in turn are greatly influenced by molecular structure and fabrication processes. By understanding the relationships between these factors, scientists are able to create novel materials for specific functions.
In the present study, researchers at NIMS applied an information science-based approach to materials research to decipher relationships between material properties, molecular structure and fabrication processes. Unlike previously available methods which rely on analyzing numerical data of materials, the CAMaD system uses natural language processing and weekly supervised deep learning to identify key material features and draw links to molecular structure and fabrication methods.
With the CAMaD system, material designers first have to select several material properties relevant to desirable material performance. Subsequently, the software automatically determines the type and strength of relationships between material structures deemed responsible for the desired material properties. Structure-controlling fabrication processes are also highlighted in the eventual report.
The use of the CAMaD system enables information from thousands of scientific and technical articles to be summarized in a single chart, rationalizing and expediting material design.
“In this pioneering effort, we actively integrated natural language processing and deep learning into material design. We have publicized the AI source code developed in this study for use by others, free of charge, to promote related research,” said the researchers.
The article can be found at: Onishi et al. (2018) Relation Extraction With Weakly Supervised Learning Based on Process-structure-property-performance Reciprocity.
———
Source: National Institute for Materials Science; Photo: Shutterstock.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
