PyMerger: Detecting Binary Black Hole Mergers from the Einstein Telescope Using Deep Learning

Abstract We present PyMerger , a Python tool for detecting binary black hole (BBH) mergers from the Einstein Telescope (ET), based on a deep residual neural network (ResNet) model. ResNet was trained on data combined from all three proposed subdetectors of ET (TSDCD) to detect BBH mergers. Five different lower-frequency cutoffs ( F low )—5 Hz, 10 Hz, 15 Hz, 20 Hz, and 30 Hz—with the match-filter signal-to-noise ratio (MSNR) ranges 4–5, 5–6, 6–7, 7–8, and >8 were employed in the data simulation. Compared to previous work that utilized data from a single subdetector, the detection accuracy from TSDCD has shown substantial improvements, increasing from 60%, 60.5%, 84.5%, 94.5%, and 98.5% to 78.5%, 84%, 99.5%, 100%, and 100% for sources with MSNRs of 4–5, 5–6, 6–7, 7–8, and >8, respectively. The ResNet model is evaluated on the first ET mock data challenge (ET-MDC1) data set, where the model demonstrates strong performance in detecting BBH mergers, identifying 5566 out of 6578 BBH events, with optimal SNRs starting from 1.2 and a minimum and maximum D L of 0.5 Gpc and 148.95 Gpc, respectively. Despite being trained only on BBH mergers without overlapping sources, the model achieves high BBH detection rates. Notably, even though the model was not trained on binary neutron star (BNS) and black hole-neutron star (BHNS) mergers, it successfully detected 11,477 BNS and 323BHNS mergers in ET-MDC1, with optimal SNRs starting from 0.2 and 1, respectively, indicating its potential for broader applicability.