Application of machine learning in ground-based gravitational wave transient noise data processing

In 2015, LIGO first detected the binary black hole merger event GW150914, marking a new era in gravitational wave astronomy. Ground-based laser interferometric gravitational wave detectors possess extremely high sensitivity, making them susceptible to transient noise (glitches) from both the instrument and the environment. Despite the development of various methods and tools in gravitational wave signal processing to enhance the characterization, identification, and removal of these glitches, LIGO's continuous data output can overwhelm any research team. Therefore, developing robust methods to manage glitches is crucial for improving the reliability and scientific value of gravitational wave detection.This article first reviews the advantages and disadvantages of general noise data processing methods, then outlines the applications of various machine learning techniques in this area. Finally, it introduces the latest application of machine learning in glitch data processing through the Gravity Spy project. The goal is to present recent advancements in machine learning for the recognition, modeling, and removal of glitches, providing a reference for researchers handling transient noise in ground-based gravitational wave data.