Searching for Precessing Black Hole Binaries in Gravitational-wave Data

In this dissertation, we described the four years long journey to detect gravitational waves emitted by precessing binary black holes systems, also called precessing signals. Our work was motivated by the apparent rarity of detection of precessing signals, which raised the obvious question of why so few signals with clear imprint of precessions have been observed. Given the large scientific interest of precessing signals, with impacts on BBH formation mechanisms and cosmology, the question is very worthy of an answer. Therefore, to rule out the possibility that a few precessing signals were missed by the lack of sensitivity of current aligned-spin searches, we developed a technique to search for precessing signals, achieving up to 120% sensitivity increase over traditional searches. Thanks to our work, we were able to search for the first time for precessing signals in a small but interesting corner of the precessing parameter space and we were able to rule out the presence of precessing signals not detected by previous searches, thus confirming the results from the LIGO, Virgo and KAGRA collaboration. This work was made possible by three key studies. The first tackled the challenge of placing search templates in a complex, high-dimensional space, using a machine learning model called a normalizing flow to efficiently sample from the required space. The second focused on improving the signal consistency check used in matched-filtering searches. The third combined these advancements into an upgraded version of the GstLAL search pipeline, which we thoroughly tested for its ability to detect precessing black hole signals.