2026-04-02 中国科学院（CAS）
Data-driven research for functional ceramics (Image by SICCAS)
＜関連情報＞
- https://english.cas.cn/newsroom/research-news/202605/t20260509_1158655.shtml
- https://www.sciencedirect.com/science/article/abs/pii/S0927796X26000380
機能性セラミックスのためのデータ駆動型アプローチ Data-driven approaches for functional ceramics
Jincheng Qin, Faqiang Zhang, Mingsheng Ma, Yongxiang Li, Zhifu Liu
Materials Science and Engineering: R: Reports Available online: 1 April 2026
DOI:https://doi.org/10.1016/j.mser.2026.101213
Abstract
Functional ceramics play a pivotal role in modern technologies due to their diverse responsive properties. However, their design and properties optimization remain challenging due to complex composition-structure-property relationships and high-dimensional parameter spaces. Data-driven approaches, particularly machine learning (ML), offer powerful tools to accelerate the discovery and development of functional ceramics by enabling rapid property prediction, knowledge discovery, and decision-making. This review summarized recent advances in ML applications for functional ceramics. Beginning by introducing the end-to-end ML workflow, including data collection, featurization, algorithm selection, model evaluation and interpretation, the progress in properties prediction were summarized across major classes of functional ceramics, including dielectric, ferroelectric, piezoelectric, electrocaloric, conductive, superconductive, magnetic, and luminescent materials, organized by their key properties. Beyond property prediction, we highlighted ML applications in materials classification, calculation enhancement, process optimization, pattern recognition, device design and failure analysis. Finally, the emerging challenges and opportunities in data standardization, intelligent experimentation, small-data learning, multimodal fusion, explainable AI, digital twin and exploration of novel ceramics were discussed. This review aims to serve as a guide and inspiration for researchers leveraging data-driven strategies to enable intelligent design and deployment of high-performance functional ceramics.
