AstroAI Lunch Talks - July 20, 2026 - Belén Costanza
20 Jul 2026 - Joshua Wing
The video can be found here: https://www.youtube.com/watch?v=QdgMd0RQDF8
Speaker: Belén Costanza (National University of La Plata)
Title: Deep Learning Applications in Cosmology: From the Early to the Late Universe
Abstract: Deep learning has become a powerful tool in modern cosmology, offering new strategies to extract and maximize cosmological information. In this talk, I will review neural network methodologies developed during my PhD, addressing two distinct regimes. For the early universe, I will present a neural network designed to perform Wiener filtering on Cosmic Microwave Background polarization maps, enabling accurate reconstruction of E- and B-mode signals while optimizing power spectrum estimation. For the late universe, I will show how deep learning can distinguish between different dark matter models using hydrodynamical simulations of large-scale structure, highlighting emerging connections between galaxy observables and the underlying nature of dark matter.