Bloch oscillations in graphene from an artificial neural network study

TL;DR

This paper introduces an artificial neural network method to classify Bloch oscillation signals in graphene, achieving high accuracy in determining the electric field orientation and strength, with potential applications to other Dirac-Weyl materials.

Contribution

The study presents a novel ANN-based approach for classifying Bloch oscillation signals in graphene, including solving the inverse problem of identifying electric field parameters.

Findings

Classification accuracy ranges from 82.6% to 99.3%.

Method can be extended to other Dirac-Weyl materials.

Performance depends on training time and computational resources.

Abstract

We develop an artificial neural network (ANN) approach to classify simulated signals corrsponding to the semi-classical description of Bloch oscillations in pristine graphene. After the ANN is properly trained, we consider the inverse problem of Bloch oscillations (BO),namely, a new signal is classified according to the external electric field strength oriented along either the zig-zag or arm-chair edges of the graphene membrane, with a correct classification that ranges from 82.6% to 99.3% depending on the accuracy of the predicted electric field. This approach can be improved depending on the time spent in training the network and the computational power available. Findings in this work can be straightforwardly extended to a variety of Dirac-Weyl materials.