Knowledge of the domain of applicability of a machine learning model is essential to ensuring accurate and reliable model predictions.In this work,we develop a new and general approach of assessing model domain and de...Knowledge of the domain of applicability of a machine learning model is essential to ensuring accurate and reliable model predictions.In this work,we develop a new and general approach of assessing model domain and demonstrate that our approach provides accurate and meaningful domain designation across multiple model types and material property data sets.Our approach assesses the distance between data in feature space using kernel density estimation,where this distance provides an effective tool for domain determination.We show that chemical groups considered unrelated based on chemical knowledge exhibit significant dissimilarities by our measure.We also show that high measures of dissimilarity are associated with poor model performance(i.e.,high residual magnitudes)and poor estimates of model uncertainty(i.e.,unreliable uncertainty estimation).Automated tools are provided to enable researchers to establish acceptable dissimilarity thresholds to identify whether new predictions of their own machine learning models are in-domain versus out-of-domain.展开更多
基金the Bridge to the Doctorate:Wisconsin Louis Stokes Alliance for Minority Participation National Science Foundation(NSF)award number HRD-1612530the University of Wisconsin-Madison Graduate Engineering Research Scholars(GERS)fellowship program,and the PPG Coating Innovation Center for financial support for the initial part of this work.The other authors gratefully acknowledge support from the NSF Collaborative Research:Framework:Machine Learning Materials Innovation Infrastructure award number 1931306+1 种基金Lane E.Schultz also acknowledges this award for support for the latter part of this work.Machine learning was performed with the computational resources provided by XSEDE 2.0:Integrating,Enabling and Enhancing National Cyberinfrastructure with Expanding Community Involvement Grant ACI-1548562We thank former and current members of the Informatics Skunkworks at the University of Wisconsin-Madison for their contributions to early aspects of this work:Angelo Cortez,Evelin Yin,Jodie Felice Ritchie,Stanley Tzeng,Avi Sharma,Linxiu Zeng,and Vidit Agrawal.
文摘Knowledge of the domain of applicability of a machine learning model is essential to ensuring accurate and reliable model predictions.In this work,we develop a new and general approach of assessing model domain and demonstrate that our approach provides accurate and meaningful domain designation across multiple model types and material property data sets.Our approach assesses the distance between data in feature space using kernel density estimation,where this distance provides an effective tool for domain determination.We show that chemical groups considered unrelated based on chemical knowledge exhibit significant dissimilarities by our measure.We also show that high measures of dissimilarity are associated with poor model performance(i.e.,high residual magnitudes)and poor estimates of model uncertainty(i.e.,unreliable uncertainty estimation).Automated tools are provided to enable researchers to establish acceptable dissimilarity thresholds to identify whether new predictions of their own machine learning models are in-domain versus out-of-domain.
