深度学习在引力波探测中的应用综述

A review of the application of deep learning in gravitational wave detection

引力波探测是现代科技史上的重要事件。随着引力波探测技术的发展,传统的匹配滤波方法逐渐不符合现代引力波探测的要求,因此寻找新的探测方法势在必行。本文回顾了引力波探测的主要研究成果,分析了深度学习方法在引力波探测中的应用潜力。在引力波信号检测方法中,引入卷积神经网络(CNN)模型的深度滤波方法获得了不低于匹配滤波的识别正确率和大幅提升的处理速度。在进行引力探测数据分析、对信号中的噪声进行识别与分类时,深度学习方法也有着优秀的表现。最后,介绍了利用卷积神经网络实现引力波识别的具体流程。

Gravitational wave detection has been an important event in the history of modern science and technology.With the fast development of gravitational wave detection technology,traditional methods such as the matched filtering method cannot meet the requirements of modern gravitational wave detection.Hence,it is necessary to develop the new detection methods for gravitational wave.This work reviews the results of major research teams on gravitational wave detection in recent years,and analyzes the application potential of deep learning methods in gravitational wave detection.Among them,the deep filtering method adopts the convolutional neural network(CNN)model.The detection accuracy of the deep filtering method is similar to results by the matched filtering method while the data processing speed is greatly improved.Besides,the deep learning method also shows excellent performance in the recognition and classification process of noise from gravitational wave signals.Finally,the specific processes of gravitational wave detection by means of convolutional neural network model are demonstrated.