Unearthing Fast Radio Bursts using Machine Learning

The thesis enhanced our understanding of Fast Radio Bursts (FRBs) by optimizing the detection efficiency of multi-beam radio interferometers. FRBs are impulsive radio signals visible to cosmological distances. The study developed a novel, machine-learning based, spatially-aware classification system for live FRB detections. It was implemented on the UTMOST-NS radio telescope in Australia. A very bright FRB with very high time, frequency and polarisation information was detected, a first for the facility. The method could be implemented at other telescopes to improve detection rates, crucial for resolving key questions about FRB populations and their progenitors.