Assessing and marginalizing over compact binary coalescence waveform systematics with RIFT

TL;DR

This paper introduces an efficient method to account for waveform modeling uncertainties in gravitational wave source parameter inference by marginalizing over different waveform models using the RIFT engine.

Contribution

It presents a novel, computationally efficient technique to marginalize over waveform systematics in gravitational wave analysis, compatible with various models including costly ones.

Findings

Method robustly marginalizes over waveform uncertainties.

Enables accurate inference with disjoint posterior models.

Increases reusability and efficiency of calculations.

Abstract

As Einstein's equations for binary compact object inspiral have only been approximately or intermittently solved by analytic or numerical methods, the models used to infer parameters of gravitational wave (GW) sources are subject to waveform modeling uncertainty. Using a simple scenario, we illustrate these differences, then introduce a very efficient technique to marginalize over waveform uncertainties, relative to a pre-specified sequence of waveform models. Being based on RIFT, a very efficient parameter inference engine, our technique can directly account for any available models, including very accurate but computationally costly waveforms. Our evidence and likelihood-based method works robustly on a point-by-point basis, enabling accurate marginalization for models with strongly disjoint posteriors while simultaneously increasing the reusability and efficiency of our intermediate calculations.