PhD Thesis Defense by Nikki Arendse

Supervisor: Jens Hjorth and Radek Wojtak

Examiners: Dr. Eleonora di Valentino (external), Dr. Philip James Marshall (external) and Steen Harle Hansen (internal)

Titel: Cosmic dissonance: Addressing tensions in modern cosmology

The persistent tension between low-redshift distance indicators and the Cosmic Microwave Background (CMB) radiation, often referred to as the ‘Hubble tension’, constitutes one of the key mysteries in present-day cosmology. The primary aim of this thesis is to investigate the aforementioned discrepancy by combining observational data with simulations and theoretical models, while drawing from the latest advances in the field of machine learning.Firstly, I will examine the tension in a broader framework, in terms of both the Hubble constant and the sound horizon, by employing observations of gravitationally lensed quasars, type Ia supernovae, Baryon Acoustic Oscillations, and the CMB. Moreover, I will investigate whether new cosmological models can resolve the combined tension and show that none of the modifications to our standard model manages to do so. These findings highlight the importance of novel independent measurements of the cosmic expansion rate.The second part of the talk explores one such new avenue to infer the Hubble constant: the use of gravitationally lensed supernovae for distance measurements. For this purpose, I will present a deep learning framework to identify lensed supernovae from optical transient surveys through both their spatial and time-variable features. I will demonstrate the improvement in classification accuracy when using time-series images instead of single-epoch observations. Finally, I will present a similar machine learning pipeline to infer the Hubble constant from simulated time-series images of lensed supernovae as expected from the Vera Rubin Observatory.