Title:A Deep Learning Model for Atomic Structures Prediction Using X-ray Absorption Spectroscopic Data

Abstract:A deep neural network (DNN) model consisting of two hidden layers was proposed for predicting the immediate environments of specific atoms based on X-ray absorption near-edge spectra (XANES). The output layer of the DNN can be adjusted to form a classifier or regressor, to predict the local and overall coordination environments, respectively. Using Li3FeO3.5 as a model system, it was demonstrated that the prediction accuracy of the DNN classifier is higher than 98%, and the predictions of the DNN regressor also showed notable agreement with the ground truth. Therefore, despite its simplicity, this DNN architecture can be expected to be generally capable of predicting the structural properties of various systems. Fine tuning of the hyperparameters, bias-variance tradeoff, and strategies to enrich the versatility of the model were also discussed.