Using machine learning to autotune chi-squared tests for gravitational wave searches

The sensitivity of gravitational wave searches is reduced by the presence non-Gaussian noise in detector data. These non-Gaussianities often match well with template waveforms used matched filter searches, and require signal-consistency tests to distinguish them from astrophysical signals. However, empirically tuning these for maximum efficacy time consuming limits complexity tests. In this work we demonstrate a framework use machine-learning techniques automatically tune We implement new $\chi^2$ test targeting large population found intermediate mass black hole binaries, training using set out paper. find that method effectively trains complex model down-weight noise, while leaving signal relatively unaffected. This improves search $\sim 11\%$ signals masses $> 300 M_\odot$. future could be any commonly matched-filter algorithms, further improving our searches.