Neural‐Network Quantum State Surrogates for Chiral Topological Superconductors

ABSTRACT

Chiral topological superconductors and Majorana‐supporting nanowires constitute key building blocks for fault‐tolerant quantum technologies, yet their modelling remains computationally demanding due to the interplay of spin–orbit coupling, Zeeman fields, and unconventional pairing. Using the pyqula framework, we compute quasiparticle spectra, Berry curvature, Chern numbers, edge density‐of‐states maps, and real‐space LDOS profiles for non‐unitary triplet superconductors and Rashba nanowires, providing quantitative signatures directly relevant to tunneling and scanning‐probe experiments. We further construct compact neural‐network quantum‐state (NQS) surrogate models that reproduce bulk invariants and nanoscale spectroscopic features while enabling orders‐of‐ magnitude faster exploration of device‐relevant parameter spaces. This hybrid BdG–ML approach offers a practical route for modeling, interpreting, and optimizing topological superconducting platforms for emerging quantum technologies.