
The LISA Global Fit: How to Listen to 10,000 Conversations at the Same Time
Neil Cornish of Montana State University explains the data analysis challenge facing LISA, the planned space-based gravitational wave observatory. Unlike ground-based detectors such as LIGO, which see a handful of loud, short events, LISA will continuously record a dense superposition of tens of thousands of overlapping signals from compact binaries, supermassive black hole mergers, and unresolved galactic noise. Cornish lays out the Global Fit approach his team has built to disentangle these overlapping sources, treating the problem as something closer to separating thousands of simultaneous conversations than detecting isolated chirps. He walks through the Bayesian and algorithmic machinery used to jointly estimate source parameters, discusses computational cost, and addresses what can and cannot be resolved with current methods. The talk was given at the Institute for Advanced Study's 2025 Prospects in Theoretical Physics program on gravitational waves, aimed at researchers already familiar with the basics of LISA and waveform modeling.