Covariance Linkage Assimilation method for Unobserved Data Exploration

This study proposes a materials search method combining a data assimilation technique based on a multivariate Gaussian distribution with Bayesian optimization. The efficiency of the optimization using this method was demonstrated using a model function. By combining Bayesian optimization with data assimilation, the maximum value of the model function was found more efficiently. A practical demonstration was also conducted by constructing a data assimilation model for the bandgap of (Sr$_{1-x_{1}-x_{2}}$La$_{x_{1}}$Na$_{x_{2}}$)(Ti$_{1-x_{1}-x_{2}}$Ga$_{x_{1}}$Ta$_{x_{2}}$)O$_{3}$. The concentration dependence of the bandgap was analyzed, and synthesis was performed with chemical compositions in the sparse region of the training data points to validate the predictions.
Comment: 7 pages, 3 figures