Efficient Massive Black Hole Binary parameter estimation for LISA using Sequential Neural Likelihood
Abstract
LISA (Laser Interferometer Space Antenna) is a space mission that has recently entered the implementation phase. The main goal of LISA is to survey, for the first time, the observable universe by detecting low-frequency gravitational waves (GWs). One of the main GW sources that LISA will detect is the coalescence and merger of Massive Black Hole Binaries (MBHBs). In this work, we explore the application of Sequential Neural Likelihood, a simulation-based inference algorithm, to the detection and parameter estimation of MBHBs signals in the LISA data. Instead of sampling from the conventional likelihood function, which typically relies on the assumptions of Gaussianity and stationarity of the noise and requires a forward simulation for each evaluation, this method constructs a surrogate likelihood that is ultimately described by a neural network trained from a dataset of simulations (noise + MBHB). Since the likelihood is independent of the priors, we can iteratively train models that target specific observations in a fraction of the time and computational cost that other traditional and machine learning-based strategies would require. One key aspect of the method is the compression of the data. Because of the iterative nature of the method, we are able to train models to obtain qualitatively similar posteriors with less than 2% of the simulator calls that Markov Chain Monte Carlo methods would require. We compare these posteriors with those obtained from standard techniques and discuss the differences that appear, some of which can be identified as caused by the loss of information from the data compression we use. We also discuss future improvements to this method.
- Publication:
-
arXiv e-prints
- Pub Date:
- June 2024
- DOI:
- 10.48550/arXiv.2406.00565
- arXiv:
- arXiv:2406.00565
- Bibcode:
- 2024arXiv240600565M
- Keywords:
-
- General Relativity and Quantum Cosmology;
- Astrophysics - High Energy Astrophysical Phenomena;
- Astrophysics - Instrumentation and Methods for Astrophysics
- E-Print:
- RevTeX 4.2, 28 pages, 17 figures, 20 plots. Comments welcome