Zafiirah Hosenie Awarded IoP CPG Thesis Prize

First prize for this year’s 2022 IoP CPG Thesis Prize has been awarded to Zafiirah Hosenie, University of Manchester. Zafiirah’s thesis, titled Feature Detection and Classification in streaming and non-streaming astronomical datasets, applied machine learning techniques to the challenges that arise from the large, streaming, data volumes that are prevalent in modern Astronomy.

Fast Radio Burst Intelligent Distinguisher (FRBID) is a machine learning model designed to filter out the so called Radio Frequency Interference (RFI) detections from true astrophysical sources (single pulses (SP) or fast radio bursts (FRB)) for real-time classification of candidates. The performance of FRBID shows a false positive rate of less than 1%. To-date, FRBID has detected more than half a dozen new single pulse candidates.

When classifying astronomical source types to their observed variations in brightness, there exists an imbalance: There are many class types that are rare but potentially quite interesting. Zafiirah enacted a rigorous statistical analysis of the features used to identify these systems, and developed novel machine learning approaches to deal with the class imbalance. Additionally, she worked with the real-time transient pipeline of the MeerLICHT telescope to resolve the problem of distinguishing between real transients and ‘bogus’ ones. Time-domain astrophysics, studying transient and variable stars, also allows astronomers to explore the Universe from a new perspective, and the algorithms Zafiirah developed have been successfully deployed at the MeerKAT radio telescope array and the MeerLICHT optical telescope, both in South Africa.

We look forward to reading more about Zafiirah’s work in the next IoP Computational Physics Group Newsletter. In the meantime, Zafiirah’s thesis is available online.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: