Machine learning study of the deformed one-dimensional topological superconductor

A one-dimensional p-wave topological superconductor deformed by a sine-square-deformation is studied in the framework of machine learning. A supervised learning algorithm is applied with a convolutional neural network to discern the existence of a Majorana zero mode, which is the hallmark of topological superconductivity. The machine learning algorithm learns features of the Majorana zero mode, and the neural network trained with the dataset from the link deformed case turns out to be the most effective.