Feedback-Conditional 3D Reconstruction of Cosmic Baryons with Flow Matching
The model learns to predict 3D cosmic gas density fields from galaxy surveys by conditioning its prediction on galaxy density, cosmological parameters, and a measure of baryonic feedback via a power spectrum suppression term. This allows for the reconstruction of per-sightline Fast Radio Burst (FRB) dispersion measures and their uncertainties. The method generalizes well to large volumes and provides an inference path to constrain astrophysical feedback models.