Abstract
Quantum transport is a stochastic process. The conductance is fully characterized by the average of electronic transmission coefficients and its fluctuations. One of the traditional methods to calculate the transmission coefficients and conductance fluctuations is the brute force calculation. A great number of configurations are needed to obtain accurate estimate for performing disorder average on various transport quantities. So, it’s time consuming for traditional methods.
Machine learning is one of the most exciting recent technologies. It was developed as a new capability for computers and today it touches many segments of industry and basic science. In my talk, I will first introduce two approaches of machine learning, fully connected neural network and convolutional neural network, and then I will present how to predict the average of transmission coefficients and conductance fluctuations by these two methods. It’s time-saving, compared with the traditional methods like brute force calculation.
Anyone interested is welcome to attend.