Title: Why a tiny neutrino particle inspired me to simulate and reconstruct the entire Universe
Abstract: Neutrinos are the only known particles whose absolute mass scale remains undetermined, and cosmology provides one of the most promising paths to measure it. In this talk, I will first explore how different components of the cosmic web can be leveraged to break parameter degeneracies and extract significant information about neutrino mass from the non-linear matter field. I will then assess the extent to which upcoming galaxy clustering, weak lensing, and CMB surveys can capture this non-linear information in practice. In turn, I will introduce the HalfDome cosmological simulations—a suite of full-sky simulations designed for joint analyses of Stage IV surveys—and discuss how they can be used to mitigate systematics, tighten parameter constraints, and serve as a testbed for machine learning applications. I will then motivate field-level inference as an optimal approach for extracting information from cosmic structure and reconstructing the initial conditions of the Universe—in particular, I will present different methods of field-level inference, ranging from differentiable forward modeling to neural networks, and highlight the potential for improving BAO constraints with DESI. I will finally showcase methods to improve the robustness, interpretability, and efficiency of simulation-based inference—outlining a vision for its role in next-generation cosmology.
For help, please contact Webmaster.