Following a binary black hole merger, the remnant undergoes a characteristic relaxation towards its final Kerr state through the radiation of quasinormal modes (QNMs): oscillations at specific complex frequencies dictated by black hole perturbation theory. Although these modes are routinely extracted from numerical relativity waveforms, reliably determining the full QNM content of the ringdown remains a significant challenge.
In this talk, I will present a new, fully Bayesian, data-driven framework for identifying QNMs, providing a robust alternative to conventional least-squares approaches. I will illustrate how this method enables the detection of QNMs in state-of-the-art Cauchy-characteristic evolution simulations, including overtones, retrograde modes, and nonlinear modes up to cubic order. I will also discuss its application to the search for late-time power-law tails.
