Deep neural networks for inverse problems in mesoscopic physics: Characterization of the disorder configuration from quantum transport properties

TL;DR

This paper introduces a machine learning method using neural networks to infer and reconstruct the disorder potential in a 2D electronic system from quantum transport data, enabling detailed characterization of disorder landscapes.

Contribution

The study demonstrates that neural networks can accurately recognize and reconstruct the disorder potential from transport properties, advancing inverse problem solutions in mesoscopic physics.

Findings

Neural networks can identify disorder configurations from transport data.

The method can reconstruct the complete disorder landscape.

High accuracy in characterizing unknown samples' disorder potential.

Abstract

We present a machine learning approach that allows to characterize the disorder potential of a two-dimensional electronic system from its quantum transport properties. Numerically simulated transport data for a large number of disorder configurations is used for the training of artificial neural networks. We show that the trained networks are able to recognize details of the disorder potential of an unknown sample from its transport properties, and that they can even reconstruct the complete potential landscape seen by the electrons.