Speaker
Description
The rapidly growing field of dark siren cosmology, driven by advances in Gravitational-Wave (GW) detection campaigns and galaxy surveys, is progressing toward independent and increasingly precise measurements of the Hubble constant. As statistical uncertainties shrink, it becomes crucial to control and eliminate emerging systematics in order to address cosmological and astrophysical challenges. An important one concerns the currently unknown probability that a compact binary merger occurs in a given host galaxy as a function of its physical properties (e.g. absolute luminosity) raised to some power, the galaxy weight.
In this work, we develop a population-based weighting scheme for host galaxies, in which the galaxy weight is treated as a population-level parameter and jointly inferred from the data within a high-dimensional space (>>100) that includes single-event, population, and cosmological parameters. We apply our framework to O5-like simulated GW events using the MICECATv2 mock galaxy catalog, and we also obtain results for real data using the single-event GW170817 together with the GLADE+ galaxy catalog.
On one hand, this approach enables a fully Bayesian ranking of the host candidates for GW events, allowing us to identify the most likely host galaxies and potentially reveal connections between their astrophysical properties and those of the hosted mergers. On the other hand, by exploiting Hamiltonian Monte Carlo algorithms, we can tackle the full hierarchical inference problem, exploring correlations across parameters at all levels, while avoiding sources of numerical systematics associated with multidimensional integrals in the total likelihood, which could soon become a limiting factor for current analysis pipelines.