Bayesian Techniques for Astrophysical Inference from Gravitational-waves of Compact Binary Coalescences: an Application to the Third LIGO-Virgo-KAGRA

LIGO-Virgo探测器网络对引力波的观测使用了贝叶斯参数估计和模型选择技术来描述信号起源处的双星黑洞和双星中子星星的特征。这些方法是引力波天体物理学新领域的核心。该项目将改进现有技术，以促进我们对下一次引力波观测的理解。特别是，探测器的噪声特性还没有得到很好的理解，可能的重点是开发噪声模型和探测器特定的策略，以便通过对探测器数据和监测系统使用数据挖掘和机器学习技术，从引力波观测中进行准确的测量。另一个研究方向是将多个引力波观测结果汇集在一起。这样，贝叶斯分析可以推断出潜在分布的参数，例如天体物理种群的共同属性，引力理论的约束，以及物质状态方程的测量。这将包括阈下事件;单独观察不那么重要，但总体上具有很大的统计功效。探测器的噪声特性对其中很大一部分很重要。

The observations of gravitational waves by the LIGO-Virgo detector network have used Bayesian parameter estimation and model selection techniques to characterises the binary-black-holes and the binary-neutron-star at the origin of the signals. Those methods are central to the new field of gravitational-wave astrophysics. This project will improve existing techniques to advance our understanding of the next gravitational-wave observations. In particular, the detector's noise properties are not well understood, and a possible focus will be to develop noise models and detector-specific strategies to enable accurate measurements from gravitational-wave observation, using data mining and machine learning techniques on both detector data and monitoring systems. Another research direction involves pooling together multiple gravitational-wave observations. That way, a Bayesian analysis can infer underlying distribution's parameters, such as the common properties of astrophysical populations, constraints of theories of gravity, and measurements of the equation-of-state of matter at its densest. This will include sub-threshold events; observations which individually are not so significant but taken in aggregate possess great statistical power. And the detectors' noise properties will be important for a very large fraction of them.