Machine learning, real-time techniques for the discovery, analysis and study of Fast Radio Bursts

Fast Radio Bursts (FRBs) are brief, bright flashes observed in the radio spectrum whose origins remain a mystery. Originating from half-way through the universe, and carrying the imprint of the media they propagate through on their billion-year-long journey, these bursts promise to further our understanding of the cosmos. This thesis describes an FRB survey performed using the Molonglo Radio Telescope. With the aid of a machine learning signal processing algorithm developed as part of this work, 6 new bursts were discovered. Noteably, and at the time of discovery, one of them was the first to be studied at the finest time resolution.