RUI: Expanding our View of the Gravitational-Wave Sky with Machine Learning and Numerical Relativity

Christopher Newport University, Newport News VA

Investigators

Abstract

This award supports research in relativity and relativistic astrophysics and it addresses the priority areas of NSF's "Windows on the Universe" Big Idea. A century after Einstein predicted the existence of gravitational waves, the two Laser Interferometer Gravitational-wave Observatory (LIGO) detectors opened a new window on the universe by discovering gravitational waves passing through Earth, emanating from cataclysmic, distant events: colliding black holes and neutron stars. The phenomenal precision that LIGO needs to clearly observe these faint waves requires exquisitely isolated detectors. Capturing the physics of colliding black holes and neutron stars requires accurate waveform models. This award will establish a new research group at Christopher Newport University (CNU) to address both of these challenges, through characterizing LIGO detectors to better understand the origins of problematic noise in the detectors, and by improving the waveform models used to interpret the astrophysics of observed signals. Through these projects, the CNU group will play a crucial role in improving the quality of the LIGO detector data and the accuracy of the parameter estimation information that is shared with the astronomical community and the public. The students supported by this award will be trained in computer programming, data analysis, and machine learning. These important transferable skills will prepare the students for a wide range of successful and meaningful STEM careers in academia and industry. LIGO’s sensitivity to gravitational waves is limited by non-stationary noise, which fluctuates over time depending on various environmental influences. This work will extend a method developed by the PI with collaborators to correlate these variations in sensitivity with auxiliary instrumental sensors to determine the most possible causes, using lasso linear regression. This method has already been useful for identifying noise sources that change over the course of hours, but this work will target problematic persistent noise transients, which impede gravitational wave searches and have rates varying over the course of days and weeks. The PI and students will also contribute to data quality validation of gravitational wave candidate events, to ensure that the broader astronomical community has access to necessary data quality information. The expected improvements to LIGO in the coming years will enable the observation of many more black holes, doubtless some with interestingly different properties and some potentially having much higher signal-to-noise ratios. Accurately extracting the astrophysical parameters of these signals requires comparing to template waveforms that span the potential discovery space. The PI and her students will work on surrogate modeling (a way to efficiently interpolate between expensive but accurate numerical relativity waveforms), working with members of the Simulating eXtreme Spacetimes collaboration. This work will enhance our ability to interpret black hole observations, especially those with large spins. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.