We develop a neural network approach to solve the self-consistent Bogoliubov-de Gennes equations in strongly disordered $s$-wave superconductors. The method accurately reproduces inhomogeneous gap distributions and generalizes to system sizes far larger than those used in training. It reduces computational scaling from $\mathcal{O}\left(N^6\right)$ to $\mathcal{O}\left(N^2\right)$, enabling quantitative analysis of percolation phenomena and the superconductor-insulator transition with much smaller computational efforts.

Close

PrevNext

View figure in article

• Impurities in superconductors
• Bogoliubov-de Gennes equations
• Data mining
• Methods in superconductivity