Gravity Seminar: Modeling compact binary waveforms using Machine learning methods

Author: Maria Teresa Mateu Luceca

Abstract: Machine learning methods are rapidly becoming more important in gravitational wave science, to distinguish signals from noise, for parameter estimation, and to model the signals across high dimensional parameter spaces. In this talk, we present our recent work developing machine learning models to fit the remnant properties of binary black hole mergers throughout the whole parameter space. These quantities will be used to improve the ringdown part of the current phenomenological waveform models. We first train an aligned spin and then a precessing deep learning model with numerical relativity (NR). We also add analytical knowledge to our training dataset in the extreme mass ratio regime and discuss the application to other quantities.

Date: 12/01/2022, 12h.

Place: Edifici Antoni Maria Alcover i Sureda, SAL04b, UIB