Screening method discovers YRu3B2 and LuRu3B2 superconducting materials, and is deployable to find thousands more
https://www.eurekalert.org/news-releases/1133828 "machine learning narrows down promising elemental combinations, pre-screens with algorithm, then calculates which materials could be superconductive... both materials gain resistance-free electrical properties from electrons forming flat bands within kagome lattice... no slow trial-and-error, instead uses high-throughput computational screening process predicting candidate structures... realizes major energy-efficiency gains for global lossless power grids and ultra-fast quantum processors" .