A method and device for identifying an exciton polaron two-dimensional electronic spectrum

Abstract

The application relates to an exciton polaron two-dimensional electron spectrum identification method and device, wherein the exciton polaron two-dimensional electron spectrum identification method comprises the following steps: obtaining a to-be-detected two-dimensional electron spectrum spectrum; inputting the to-be-detected two-dimensional electron spectrum spectrum into a target neural network model; wherein the target neural network model is obtained based on a pre-constructed compound loss function, and the compound loss function comprises a physical constraint loss term; the physical constraint loss term is constructed based on a quantum mechanics analytical relationship of an exciton polaron system; the to-be-detected two-dimensional electron spectrum spectrum is predicted for a preset number of times through the target neural network model, a prediction result corresponding to each prediction is output, a prediction result set is obtained, and an identification result of the to-be-detected two-dimensional electron spectrum spectrum is determined based on the prediction result set. Through the application, the problems of low artificial spectrum solving efficiency and possible violation of basic physical laws of pure data-driven model prediction results are solved.

Images