AI-Enabled Discovery and Characterization of Topological Superconductors

A central challenge in realizing topological superconductivity (TSC) is the scarcity of experimentally confirmed materials, largely due to the intricate interplay of crystalline symmetry, electronic correlations, and topology. I will present an AI-based framework that automatically identifies candidate TSCs and determines the topological nature of their superconducting phases directly from ab initio band information. By combining momentum-space and real-space classification theories, this AI-enabled "TSC Materials Identifier" not only predict whether superconductors with a certain crystal structure can host Majorana modes, but also specifies their expected boundary type—edge, surface, hinge, and/or corner. The tool thus offers a fast, interpretable way to screen materials and guide experiments in real time, bridging theory and experiment in the search for new topological superconductors.