Generating High-Fidelity HI Maps Using Score-Based Diffusion Models - Sultan Hassan

Explore the application of score-based diffusion models for generating high-fidelity HI maps in this 27-minute conference talk by Sultan Hassan from NYU. Delve into the intersection of astrostatistics, machine learning, and galaxy formation physics. Discover how these advanced techniques can be applied to large-scale radiative transfer, non-Gaussian generative models, and invertible mapping. Learn about the implications for cosmic microwave background studies and the potential for enhancing our understanding of galaxy evolution. Gain insights into the latest developments in data-driven tools for exploring galaxy formation physics and their potential to maximize information extraction from current and future astronomical surveys.

Introduction
Largescale radiative transfer
Nongaussiangenerative models
Invertible mapping
Diffusion models
Cosmic microwave background
Results
Summary