Machine learning for gravitational waves

The discovery of gravitational waves by the LIGO/VIRGO collaboration has been a paradigm shift in modern astrophysics. It confirmed the predictions Einstein’s General theory of Relativity and opened a new window into the Universe and fundamental physics [1]. Current observations of neutron star and black hole mergers will be complemented with the wealth of future ground- and space-interferometers, predominantly the Einstein Telescope and LISA [2]. Translating the information in gravitational waveforms to implications for fundamental physics is a computationally demanding. This is where machine learning tools are of great help [3, 4].

This project will develop advanced machine learning techniques for the detection and fast parameter inference of gravitational waveforms from astrophysical systems, towards probing theories of fundamental physics. We will use state-of-the-art deep algorithms in Python in conjunction with parallel/GPU computing methods for the simulation of waveforms and the training of deep networks. Part of the research will be embedded within the international consortia of LISA and Einstein Telescope, and the student will have the unique opportunity to interact with a broad range of scientists – from theorists and phenomenologists to software engineers.

Expected: Advanced coding skills in Python or a clear and demonstrable potential to develop them is essential. Prior experience with machine learning and/or the theory or modelling of gravitational waves is a plus, but not necessarily required.

References:

[1] I. D. Saltas et al (2018) — arXiv: 1812.03969
[2] E. Barausse et al (2020) — arXiv: 2001.09793
[3] E. Cuoco et al (202) — arXiv: 2005.03745
[4] G. Ventagli & I. D. Saltas (2024) — arXiv: 2405.17908