GWKokab: An Implementations to Identify the Properties of Multiple Population of Gravitational Wave Sources

The increasing sensitivity of gravitational wave detectors is enabling the detection of an ever-greater number of binary mergers. These events are crucial for understanding the population properties of compact binaries. However, many previous studies have used inference implementations with prohibitively high computational costs to infer population properties. In this work, we present GWKokab, a JAX-based framework designed to facilitate the complex model building from simple components with independent rates for each subpopulation. GWKokab employs state-of-the-art Bayesian statistical methods such as normalizing flows which enables rapid inference for multiple populations. To check the robustness of the code, we generated eccentric and circular sub-populations using multi-source model, then recovered their injected parameters at significantly reduced computational costs. We also demonstrate selected science applications, including reproducing some of the results presented in two previously published studies: eccentric injections recovery and mass distribution using third Gravitational-Wave Transient Catalog (GWTC-3). Additionally, we have investigated how inferences about the BBH population would change if GWTC-3 also included a synthetic sub-population of high mass ratio binaries. We have made the code publicly available, and we anticipate ease in the subpopulation analysis of the upcoming gravitational wave events at reduced computational cost.