Precise tuning of dielectric constants(Er)in oxide glasses is critical for highfrequency devices in 5G/6G systems,whereεp directly governs signal propagation efficiency.A machine learning framework combining data aug...Precise tuning of dielectric constants(Er)in oxide glasses is critical for highfrequency devices in 5G/6G systems,whereεp directly governs signal propagation efficiency.A machine learning framework combining data augmentation and physicochemical descriptor integration is developed to address data scarcity.Validated pseudo-labels are generated via ensemble learning,expanding the dataset from 1503 to 11,029 compositions without distributional shif.The XGBoost model trained on the augmented dataset achieved superior accuracy,with an R of 0.96 and an MSE of 0.14.For prediction tasks on unseen data,it reduced the error rate by 48%compared to the non-augmented model and improved generalization performance by 43%over GlassNet.B_(2)O_(3)and SiO_(2)are identified as E,suppressors and BaO and TiO_(2)as enhancers through SHAP analysis,aligning with network former/modifier roles.Cation-specific polarizabilities are derived via ClausiusMossotti regression(R^(2)=0.909).Integration of physicochemicaldescriptors(coordination number and bond strength)enables transferable predictionsfor Y_(2)O_(3)and La_(2)O_(3)containing glasses,with mean deviation 2.46%-4.76%.Crucially,structural descriptors dominate polarizability with 69.9%feature importance,establishing network engineering as the optimal design paradigm.A data-driven pathway for rational dielectric glass development is thus established.展开更多
基金the National Natural Science Foundation of China(No.52172019)Shandong Provincial Youth Innovation Team Development Plan of Colleges and Universities(No.2022KJ100).
文摘Precise tuning of dielectric constants(Er)in oxide glasses is critical for highfrequency devices in 5G/6G systems,whereεp directly governs signal propagation efficiency.A machine learning framework combining data augmentation and physicochemical descriptor integration is developed to address data scarcity.Validated pseudo-labels are generated via ensemble learning,expanding the dataset from 1503 to 11,029 compositions without distributional shif.The XGBoost model trained on the augmented dataset achieved superior accuracy,with an R of 0.96 and an MSE of 0.14.For prediction tasks on unseen data,it reduced the error rate by 48%compared to the non-augmented model and improved generalization performance by 43%over GlassNet.B_(2)O_(3)and SiO_(2)are identified as E,suppressors and BaO and TiO_(2)as enhancers through SHAP analysis,aligning with network former/modifier roles.Cation-specific polarizabilities are derived via ClausiusMossotti regression(R^(2)=0.909).Integration of physicochemicaldescriptors(coordination number and bond strength)enables transferable predictionsfor Y_(2)O_(3)and La_(2)O_(3)containing glasses,with mean deviation 2.46%-4.76%.Crucially,structural descriptors dominate polarizability with 69.9%feature importance,establishing network engineering as the optimal design paradigm.A data-driven pathway for rational dielectric glass development is thus established.
