Recent papers involving U²GRC researchers explore complementary ways to improve gravitational-wave searches and models of merging black holes. This roundup highlights three short-author studies released in September and October 2026.
Efficient template placement
Alexandra Wernersson, Jessica Irwin, Melissa Lopez, and Sarah Caudill present Learning Approximate Isometric Embeddings for Efficient Template Placement. Their isobank method learns a coordinate transformation that makes placing waveform templates less computationally demanding. In tests on an eccentric-binary parameter space, it achieved higher coverage with fewer templates than the comparison method. arXiv | INSPIRE.
Narrowing the search for binary black holes
In Neural Network Guided Parameter Space Constraints for Gravitational Wave Searches from Binary Black Holes, Chetan Verma, Amit Reza, Gurudatt Gaur, Dilip Krishnaswamy, and Sarah Caudill extend a neural-network approach to binaries with aligned spins. The study examines how the choice of parameter coordinates affects identification of a smaller region for a subsequent matched-filter search. The results offer a route toward reducing search costs, with performance evaluated using modeled signals and Advanced LIGO design sensitivity. arXiv | INSPIRE.
More accurate precessing merger waveforms
Abhishek Ravishankar, Vijay Varma, Scott E. Field, and collaborators introduce NRSur7dq4v2: A multi-domain precessing surrogate model with improved accuracy. The model treats overlapping portions of the signal separately to improve control of waveform errors, especially during merger and ringdown. Tests show improved estimates of the remnant black hole’s mass and spin. The work also expands the modeled waveform modes and improves evaluation speed for gravitational-wave inference. arXiv | INSPIRE.
