首页|New Machine Learning Findings from Friedrich-Schiller-University Jena Described (Enhancing Glass Property Predictions Through Ab Initio-derived Descriptors)
New Machine Learning Findings from Friedrich-Schiller-University Jena Described (Enhancing Glass Property Predictions Through Ab Initio-derived Descriptors)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Research findings on Machine Learning are discussed in a new report. According to news reporting originating from Jena , Germany, by NewsRx correspondents, research stated, "The performance of ab ini tio descriptors derived from density functional theory simulations is systematic ally investigated in comparison to traditional compositional descriptors for the ability to predict glass properties utilizing machine learning algorithms. Two datasets are used for this purpose: an extensive, publicly available database in volving a wide range of oxide glasses, and a small in-house dataset covering a b roader collection of inorganic glasses from metallic to non-metallic materials." Financial supporters for this research include Federal Ministry of Education & Research (BMBF), Federal Ministry of Education & Research (BMBF), Carl Zeiss Foundation, German Research Foundation (DFG). Our news editors obtained a quote from the research from Friedrich-Schiller-Univ ersity Jena, "For the larger dataset, it was demonstrated that ab initio descrip tors offer a substantial reduction in input dimensionality while retaining nearl y equivalent predictive performance when compared to the compositional descripto rs. The combination of ab initio and compositional descriptors showed an improve ment in prediction accuracy. For the smaller dataset, the ab initio-derived desc riptors performed significantly better than the compositional descriptors, provi ding a valuable tool to improve glass property prediction in settings where the availability of data is limited. Furthermore, ab initio-derived descriptors are not only computationally inexpensive and allow extrapolation beyond the training composition space but also facilitate model interpretation. Simple yet effectiv e descriptors: ab initio-derived descriptors improve the performance of machine learning models for predicting glass properties in small datasets. They provide dimensionality reduction and enable predictions beyond the initial compositional space of the training set."
NETL
NSTL
