| 小数据机器学习算法在材料科学领域中的研究进展 |
| Research Progress of Small Data Machine Learning Algorithms in Materials Science Field |
| 投稿时间：2024-11-19 修订日期：2024-11-19 |
| DOI：10.12136/j.issn.1000-890X.2026.04.0309 |
| 中文关键词: 机器学习 小数据学习 学习策略 主动学习 迁移学习 |
| 英文关键词: machine learning small data learning learning strategy active learning transfer learning |
| 基金项目:国家自然科学基金NSAF基金重点项目（U2030203） |
|
| 摘要点击次数: 203 |
| 全文下载次数: 46 |
| 中文摘要: |
| 依靠传统试验-试错方法已无法满足当前对于高性能材料的快速研发及定向设计需要，机器学习作为人工智能的一个重要分支，通过在输入数据与目标性能之间建立映射关系，为材料的设计提供有效指导，从而缩短材料的开发周期。结合前沿研究成果，总结出材料科学领域中机器学习的构建方法；针对数据少、数据集规模小、样本不平衡等问题，材料科学领域中数据扩充的方法主要为从出版物中提取数据、构建数据库、进行数据源层面的高通量试验和计算；小数据机器学习策略主要为主动学习、迁移学习和叶贝斯优化。机器学习在材料科学领域中机遇与挑战并存。 |
| 英文摘要: |
| Relying on the traditional experiment-trial and error method could not longer meet the current demands for rapid reasearch and directional design of high-performance materials.Machine learning，as an important branch of artificial intelligence，provides effective guidance for material design by establishing a mapping relationship between input data and target performance，thereby shortening the cycle of material development.Based cutting-edge research results，the construction methods of machine learning in the materials science field are summarized.In response to issues such as small data volume，small dataset size and sample imbalance，etc，the methods for expanding datum in the materials science field mainly are extracting datum from publications，building data databases and conducting high-throughput experiments and caculations at the data source level.The main strategies for small data machine learning are active learning，transfer learning and Bayesian optimization.The opportunities and challenges of machine learning in the materials science field coexist. |
|
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
