The sensitivity of gravitational-wave detectors can be affected by sporadic short-duration noises (so-called glitches) from distinct non-astrophysical sources. Even worse, they could produce false alarms of gravitational-wave detections. Therefore, to mitigate these glitches, it is essential to find their origin. This talk explores the machine learning techniques employed to identify acoustic noise that produces glitches in the output of gravitational-wave detectors. We summarize feature extraction procedures, unsupervised and supervised learning combined with high-throughput computing to achieve this task.
Copyright. All Rights Reserved.