Over the past 6 years, gravitational wave astronomy has become a reality, as a result of the observations made by the LIGO and Virgo laser interferometers. These observations have already had a profound impact on our understanding of the astrophysics of compact objects, on cosmology and on fundamental physics. All these scientific results rely on first obtaining posterior probability distributions for the parameters of the individual events. To date, such posterior distributions have been obtained using standard sampling methods which are slow and computationally expensive. Future detectors, such as LISA, the Cosmic Explorer and the Einstein Telescope, will see orders of magnitude more events, which will require much faster tools for finding and robustly characterising all of the events in the data. In this talk, I will describe some of the challenges that we will face in the analysis of data from future detectors, and some of the machine learning approaches that are being developed to solve them. I will focus in particular on DINGO, a neural network approach that uses normalising flows to rapidly obtain samples from posterior distributions for the parameters of gravitational wave sources. This approach generates posterior distributions that are indistinguishable from those obtained using standard techniques, but in a small fraction of the time.
Jonathan Gair leads a research group on "Data Analysis” in Max Planck Institute for Gravitational Physics. His research is in the development and application of new methodology for gravitational wave data analysis and science exploitation. He is playing a leading role within the LIGO/Virgo collaboration in deriving constraints on cosmological parameters, such as the local expansion rate of the Universe (the Hubble constant), from gravitational wave observations. He is also heavily involved in the development of data analysis tools for, and assessing the potential scientific impact of, the planned ESA-led space-based gravitational wave detector LISA and currently chairs the LISA Science Group which oversees that activity. Jonathan Gair also has completed research projects connected to gravitational wave detection with pulsar timing and astrometric measurements, and on the development of computationally efficient techniques for parameter inference.
The A3D3 Seminar is a monthly lecture series that hosts scholars working across applied areas of artificial intelligence, such as hardware algorithm co-development, high energy physics, multi-messenger astrophysics, and neuroscience. Our presenters come from all four domain fields and include occasional external speakers beyond the A3D3 science areas, governmental agencies and industry. The seminar will be recorded and published in YouTube. To receive future event update, subscribe here.