Towards a computationally-efficient follow-up pipeline for blind continuous gravitational-wave searches

The sensitivity of continuous gravitational-wave (CW) searches for unknown neutron stars (NSs) is limited by their parameter space breadth. To fit within reasonable computing budgets, hierarchical schemes are used to identify interesting candidates using affordable methods. The resulting sensitivity critically depends on the number of candidates selected to follow-up. In this work, we present a novel framework to evaluate the effectiveness of stochastic CW follow-ups. Our results in realistic datasets allow for a ten-fold reduction of the computing cost of pyfstat, a well-established follow-up method. We also simplify the setup of multi-stage follow-ups by removing the need for parameter-space metrics. The study was conducted on Gaussian and real O3 Advanced LIGO data using realistic durations, and covered both isolated and binary sources. These results will have a positive impact on the sensitivity of all-sky searches in the forthcoming observing runs of the LIGO-Virgo-KAGRA collaboration.