Speaker
Description
The spatial morphology of Reionization contains information that is largely compressed away by global measurements of the volume-averaged neutral fraction. Ionized bubbles, neutral islands, and their connection to the high-redshift galaxy distribution encode the topology of reionization and the nature of the ionizing sources. In this talk, I will present TORRCH (TOmographic Reconstruction of the Reionization of Cosmic Hydrogen), a deep-learning framework for reconstructing the neutral-hydrogen fraction field from the three-dimensional distribution of Lyman-alpha emitters and non-Lyman-alpha-selected galaxies.
Using hydrodynamical simulations post-processed with radiative transfer, we construct mock galaxy surveys spanning different reionization histories and source prescriptions. TORRCH uses a 3D U-Net to learn the mapping between sparse galaxy tracers and the underlying neutral-fraction field, with Lyman-alpha emitters acting as ionization-sensitive probes and continuum-selected galaxies providing complementary information on the density field. I will show that the resulting maps recover the large-scale ionization morphology, reproduce key one- and two-point statistics of the projected neutral-fraction field, and capture the expected galaxy–IGM cross-correlation.
I will also discuss ongoing extensions toward probabilistic reionization tomography, where diffusion-based generative models can sample plausible ionization fields conditioned on observed galaxy tracers. This provides a route toward uncertainty-aware reconstructions of reionization morphology in the JWST, Roman, PFS, and SKA era.