Extends neural-network analysis of extreme-mass-ratio inspirals from detection to signal extraction and estimation of intrinsic source properties. The study explores whether learned estimates can narrow the parameter space for subsequent Bayesian …
Explores flow-matching posterior estimation for extreme-mass-ratio inspirals with a high-dimensional source parameter space. The study tests continuous normalizing flows as a faster route toward candidate posterior distributions for these long and …
Combines Q-transform representations with a compact convolutional network to detect extreme-mass-ratio inspirals. Tests on simulated space-detector data include time-delay interferometry and examine classification performance over different …