DINGO-BNS

"Real-time inference for binary neutron star mergers using machine learning" (Nature 639, 2025): neural posterior estimation from the Schölkopf and Macke groups characterizes neutron-star mergers in about one second instead of an hour, fast enough to work with live LIGO data — a flagship result for simulation-based inference as a scientific method, built on the group's sbi toolkit line.