Scholarlink

<p>Chalcogenide glasses, renowned for their exceptional optoelectronic properties, have become the material of choice for numerous microelectronic and optical devices. With the rapid development of artificial intelligence technologies in the field of materials science, researchers have increasingly incorporated machine learning (ML) methods to accelerate the exploration of composition–structure–property relationships in chalcogenide glasses. However, traditional ML methods, which predominantly rely on single-property modelling, are often inadequate for addressing the practical demand for multiproperty collaborative optimization. To overcome this limitation, this study proposes a graph-based deep learning approach and develops a novel model for the efficient prediction of key properties of chalcogenide glasses. Furthermore, relevant data were collected from the publicly available SciGlass database to construct an experimental dataset, and the performance of the developed model was systematically evaluated. The experimental results demonstrate the model’s remarkable stability and predictive performance, highlighting its potential application value in the design and development of chalcogenide glasses.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep learning-assisted attribute prediction of chalcogenide glasses based on graph classification

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月3日

查看全文
  • Hui Li,
  • Pan Liu,
  • Shaoyun Liu,
  • Bin Yang

摘要

Chalcogenide glasses, renowned for their exceptional optoelectronic properties, have become the material of choice for numerous microelectronic and optical devices. With the rapid development of artificial intelligence technologies in the field of materials science, researchers have increasingly incorporated machine learning (ML) methods to accelerate the exploration of composition–structure–property relationships in chalcogenide glasses. However, traditional ML methods, which predominantly rely on single-property modelling, are often inadequate for addressing the practical demand for multiproperty collaborative optimization. To overcome this limitation, this study proposes a graph-based deep learning approach and develops a novel model for the efficient prediction of key properties of chalcogenide glasses. Furthermore, relevant data were collected from the publicly available SciGlass database to construct an experimental dataset, and the performance of the developed model was systematically evaluated. The experimental results demonstrate the model’s remarkable stability and predictive performance, highlighting its potential application value in the design and development of chalcogenide glasses.