Artificial Intelligence to detect gravitational waves from a network of detectors

The application of artificial intelligence in astronomical data analysis is becoming more popular among the scientific community. In our recent work [1,2], we have shown the effectiveness of Particle Swarm Optimization (PSO), a well-known algorithm in the field of swarm intelligence, in addressing signal detection and parameter estimation problem related to gravitational waves from compact binary coalescences (CBCs). The fully coherent network analysis of data from multiple gravitational wave (GW) detectors is computationally expensive since it is associated with a high dimensional numerical optimization problem. In our previous work, we showed using a non-spinning 2.0 post-Newtonian order waveform and four gravitational wave detectors (two LIGO detectors, Virgo and Kagra) that PSO can achieve the same performance as a grid search with less than 200,000 templates for a component mass range of 1.0 to 10.0 solar masses at a network signal to noise ratio of 9. Currently, we are increasing the dimensionality of the optimization problem by adding spin parameters to the waveform and exploring the effectiveness of the algorithm.

[1] TS Weerathunga, SD Mohanty, Phys. Rev. D 95 (12), 2017
[2] ME Normandin, SD Mohanty, TS Weerathunga, Phys. Rev. D 98 (04), 2018