Compact representations to detect gravitational waves from extreme-mass ratio inspirals

Project Summary

Extreme Mass Ratio Inspirals (EMRIs), detectable only by the future space-based detector LISA, provide a unique opportunity to probe the environment in galactic centers and test general relativity (GR). We develop a Bayesian framework for model selection to distinguish between vacuum GR and potential deviations caused by environmental effects or modifications to GR. The challenge lies in the complexity of EMRI waveforms, so we explore novel techniques to accelerate the likelihood calculation. Starting with simple linear methods like Principal Component Analysis (PCA), we extend to machine-learning approaches such as autoencoders to compress waveforms and enhance the analysis.

Quentin Baghi
baghi@apc.in2p3.fr

• Assistant Professor at Université Paris Cité

Lukas Arda
lukas.arda@gmail.com
Natalia Korsakova