SemiVoice

electronicsforu

01/02/2025, 05:46 AM UTC

利用有限数据预测材料特性Predicting Material Properties Using Limited Data

➀ 印度理工学院（IISc）和伦敦大学学院（UCL）的研究人员开发了一种利用迁移学习预测材料特性的机器学习方法；➁ 该方法特别适用于半导体和储能等应用；➂ 研究采用了图神经网络（GNNs）和多属性预训练（MPT）框架来提高预测能力。

➀ Researchers from IISc and UCL have developed a machine learning method using transfer learning to predict material properties; ➁ The method is particularly useful for applications like semiconductors and energy storage; ➂ The study employs Graph Neural Networks (GNNs) and the Multi-property Pre-Training (MPT) framework to enhance predictive power.

---

本文由大语言模型（LLM）生成，旨在为读者提供半导体新闻内容的知识扩展（Beta）。