Skip navigation
Please use this identifier to cite or link to this item:
Title: Learning Novel Physics from Astronomical and Cosmological Observations
Authors: Kreisch, Christina Danielle
Advisors: Spergel, David N.
Contributors: Astrophysical Sciences Department
Keywords: cosmic microwave background
large scale structure
machine learning
Subjects: Astrophysics
Particle physics
Issue Date: 2021
Publisher: Princeton, NJ : Princeton University
Abstract: Spanning the Universe's history from its first moments to the present, this dissertation covers most of the Universe's history and the bulk of its constituents. With the most basic properties of neutrinos still not well understood, the neutrino sector may harbor yet undiscovered physics beyond the standard model of particle physics. In Chapter 2 we develop a new physics model in which neutrinos with mass self-interact in the early universe. Constraining this new physics with cosmic microwave background (CMB) and large scale structure (LSS) measurements, we find a novel cosmological model with preference for strong neutrino self-interactions that is statistically distinct from the standard cosmological model, ΛCDM, and fits observations well. Beyond offering a good fit to CMB and LSS measurements, this novel model offers a potential solution for simultaneously solving the Hubble constant and matter clustering tensions. In Chapter 3 we further investigate the data components driving this preference and find that stringent small-scale CMB measurements should be able to either destroy or prefer this novel model. In Chapter 4 we update constraints on the novel self-interacting neutrino model with ACT CMB measurements, which possess superior small-scale measurements to previous experiments. Intriguingly, we find stronger preference for the novel cosmological model than before, primarily driven by CMB polarization measurements at moderate scales, rather than small scales. The preference and longevity of this radically different cosmological model sets a precedent for thoroughly exploring new models to explain the evolution of the Universe. Beyond exploring physics beyond the standard model, it is imperative to fervently investigate the mysterious vast swaths of nothingness, cosmic voids. As the largest objects in our Universe, these giants can span over 100 million light years across. Despite their massive size, they remain relatively unexplored and underutilized. In Chapter 5 we explain the utility of cosmic voids in constraining cosmological models. In Chapter 6 we illustrate the sensitivity of galaxies within voids to the sum of neutrino masses. We further the exploration of how void properties are sensitive to the sum of neutrino masses in Chapter 7. In Chapter 8 we discuss the machine learning simulations used to build the GIGANTES void simulation suite, presented in Chapter 9, which contains over 1 billion cosmic voids and over 20 terabytes of data. With these simulations, we prove that voids offer unique cosmological information to galaxies. We show that cosmic voids should yield an independent measurement of the sum of neutrino mass from future large scale surveys like DESI, Euclid, SPHEREx, and the Roman Space Telescope. Finally, in Chapter 10 we leverage the GIGANTES suite to develop a new interpretable pooling operation for leveraging cosmic voids to constrain cosmological parameters. The Universe hosts mysteries from the smallest to the largest scales, and this work aims to begin unraveling them.
Alternate format: The Mudd Manuscript Library retains one bound copy of each dissertation. Search for these copies in the library's main catalog:
Type of Material: Academic dissertations (Ph.D.)
Language: en
Appears in Collections:Astrophysical Sciences

Files in This Item:
File Description SizeFormat 
Kreisch_princeton_0181D_13883.pdf34.43 MBAdobe PDFView/Download

Items in Dataspace are protected by copyright, with all rights reserved, unless otherwise indicated.