Abstract: Gravitational-wave observations provide a unique probe of compact objects and their environments, offering insights into stellar evolution, dynamics, cosmology, nuclear physics, and gravity. Proposed formation channels of merging binary black holes and neutron stars, such as isolated binaries, triples, dense stellar clusters, or active galactic nucleus disks, make distinct predictions on subtle signatures including spin orientations, the abundance of massive or eccentric mergers, and the possibility of local gravitational lensing. As gravitational-wave detectors improve and catalogs grow, extracting these signatures requires increasingly precise inference techniques.
In this seminar, I will discuss recent efforts to improve our ability to test binary black hole formation scenarios with gravitational wave observations. I will present a new line of evidence for spin-orbit misalignment in the observed population. I will also describe inference and search techniques that incorporate gravitational wave radiation modes beyond the quadrupole, enhancing sensitivity to massive black hole mergers and improving low-latency localization of neutron-star--black-hole mergers, essential for multimessenger astronomy. Finally, I will discuss ongoing work aimed at identifying gravitational lensing signatures from single events and the role of machine-learning-based inference in meeting the computational challenges posed by future observing runs and next-generation detectors.