利用深度学习进行致密双星并合引力波数据处理的研究

第二代地面激光干涉引力波探测器已多次捕捉到来自致密双星系统并合的引力波，这说明现有的双星引力辐射理论模型和基于匹配滤波的数据处理流程是可靠的。然而随着引力波事件的不断涌现，现有的数据处理流程会带来较大的时间延迟，不利于进一步的天文观测。本项目拟利用深度神经网络结构，建立观测致密双星并合引力辐射的深度学习框架。重点研究深度学习用于致密双星系统并合过程引力波探测的高效性和参数估计的准确性，实现致密双星系统并合引力波智能观测的目标。本项目的研究成果对促进多信使天文学的发展具有重要的科学意义，有助于实现对其它类型引力波源的智能化观测。

The Advanced Laser Interferometer Gravitational-wave Observatories (Advanced LIGO and Virgo) have repeatedly captured gravitational wave (GW) signal from compact binary coalescences (CBC), which indicates the current theoretical model for compact binary system GW radiation and data analysis pipeline based on matched filtering are reliable. However, with the continuous emergence of GW events, the current data processing will result in high latency, which is not conducive for further astronomical observation. In this project, we intend to use deep neural network structures to establish a deep learning framework for observing GW from compact binary coalescences. We will focus on the efficiency and accuracy of deep learning for the GW detection from compact binary coalescences, and achieve the goal of intelligent detection for CBC GW. The research results have important scientific significance to promote the development of multi-messenger astronomy, and help to realize the intelligent observation of GW from other sources.