A. GW Discovery and Early Warning
Project
During LIGO’s third observing run, the LIGO-Virgo Scientific Collaboration (LSC) have issued alerts about gravitational wave signals they have detected to the astronomy community in close to real time. Only five groups around the world take part in this ‘online’ search, where data from the LIGO and Virgo instruments is processed on-the-fly and alerts are issued within seconds of the data being processed.
One of the online search pipelines is the SPIIR (Summed Parallel Infinite Impulse Response) pipeline, developed and operated by scientists at UWA. SPIIR is one of the fastest of the five online pipelines. As the LSC prepares for O4, the SPIIR team is focussing on sending alerts about gravitational waves up to 30 seconds before objects like binary neutron stars finally merge. This will enable astronomers to catch neutron star mergers as they happen, rather than several hours later, and will reveal important insights about the physics of these objects.
As a student working with the SPIIR group, you will use the pipeline to learn more about the astrophysics of gravitational wave sources, and you will be responsible for running the pipeline in its ‘online’ configuration to send alerts about gravitational waves to the astronomy community, and using the pipeline to characterize the properties of merging compact objects.
Eligibility
Applicants should have excellent academic records and an undergraduate degree in physics or astronomy. General UWA PhD entrance requirements can be found on the Future Students website.
Suggested reading
LIGO website
GCN Notices
Hooper et al. 2012
Dr Qi Chu’s PhD thesis
Dr Shaun Hooper’s PhD thesis
B. GW-EM Inference
Project
As we observe more and more compact object mergers (binary black hole mergers, binary neutron star mergers and neutron star-black hole mergers), we can begin to use statistics to probe many of the properties of the population of these objects.
Just one such key question we can begin to ask: ‘Is the population of binary black hole mergers distributed isotropically in the Universe? Similar analysis of the distribution of Gamma-Ray bursts uncovered their cosmological origin and answers many important questions about the origin of these events, including their connection to host galaxies, the properties of their host galaxies and the different mechanisms that power short and long GRBs.
In this project, you will use Bayesian inference and develop methods to find answers to questions like ‘Are binary black holes distributed isotropically in the Universe?
Project
In 2017, a binary neutron star merger detected via gravitational waves was also observed across the electromagnetic spectrum. This event, known as GW170817, opened up many new possibilities for understanding the astrophysics of compact objects.
With a large number of events observed during LIGO’s 3rd observing run, it is now possible to search for any connection between these observed events, and electromagnetic transients. Of particular interest are the mysterious ‘fast radio bursts’, which at present have an unknown astrophysical origin.
In this project, you will use gravitational wave data and observations from radio and gamma-ray observatories to investigate the putative connection between short bursts of gravitational waves observed by LIGO and Virgo, fast radio bursts and gamma-ray bursts, both in archival data and in real time during the O4 observing run.
Eligibility
Applicants should have excellent academic records and an undergraduate degree in physics or astronomy. General UWA PhD entrance requirements can be found on the Future Students website.
Suggested reading
LIGO website
Stiskalek et al. 2020
James et al. 2019
C. Machine Learning and GPU Acceleration
Project
Many areas of scientific research are now leveraging computational techniques such as machine learning to solve complex problems. Gravitational wave astronomy is no different, and many areas of gravitational wave discovery include problems that are well-suited to application of machine learning techniques.
Discovery of gravitational waves in time-series data is one such problem. Identification of peaks in signal-to-noise ratio time series and distinguishing between peaks that result from true gravitational wave signals and peaks that result from noise or glitches in the detectors is a problem ideally suited to machine learning applications, where real and simulated detector data can be used to train and verify the chosen algorithms.
Another problem suited to machine learning solutions is the localization of the source of gravitational waves. Accurate localization is a vital part of gravitational wave astronomy as it enables the rapid followup to detect electromagnetic counterparts, such as the kilonova associated with the neutron star merger GW170817. Neural networks can be used to rapidly produce accurate localizations of GW mergers.
Modern computer architecture can be used to reduce the latency (time between receiving GW data and the end of data processing to send alerts of GW events) of pipelines like SPIIR. GPU acceleration has already been used to improve the latency of SPIIR, and further refinement of the algorithms used in the pipeline can further reduce the pipeline’s latency. What’s more, many machine learning algorithms are well suited to implementation on GPU hardware, an ideal project for students interested in both optimization and machine learning.
In this project, you will use machine learning techniques and GPU acceleration to develop new and improved ways to detect and localize GWs in real time.
Eligibility
Applicants should have excellent academic records and an undergraduate degree in physics or astronomy is preferred. Applicants should have skills in programming or an interest in developing their programming skills. General UWA PhD entrance requirements can be found on the Future Students website.
Suggested reading
LIGO website
Chatterjee et al. 2019
Cuoco et al. 2020
For further information on any of the above projects, contact Prof. Linqing WEN [email protected] and Ruby Chan [email protected]