The current working title of the thesis is "Bayesian Methods and Stochastic Variability Modelling in High-Energy Astrophysics".

本论文的目的是回顾、应用和扩展现有的天文数据贝叶斯统计建模方法。特别是，这项研究的主要重点是对类星体或X射线双星等高能天体物理源的时间变异性进行建模。在这一主题下，该项目的结构将结合两个主要的研究分部分。第一个分部分将试图解决天文信号源重叠的问题，目前正在开发的方法在统计卫星图像处理方面有潜在的应用。天体物理学界担心，SpaceX卫星星座将污染备受期待的维拉·C·鲁宾天文台的遗产时空调查(LSST)产生的图像。这种方法可以用来对LSST图像中的卫星轨迹进行预处理，并为后续分析提供清晰的图像。除了改进和扩展目前用于解决这一问题的模型外，该项目还专注于设计高效和可扩展的算法，并开发一种强大的工具，使天体物理学家可以整合到他们目前的研究过程中。第二个子组件将试图通过提供感兴趣的物理量的概率估计来解决哈勃常数测量不一致的问题。这包括开发灵活和可扩展的方法来对天文时间序列进行建模。研究方法的新颖性在于设计了最先进的建模和计算技术，以从复杂的数据中获得新的见解。这项研究的直接影响将是对天文数据的统计分析，但所开发的方法被认为具有高度的普适性，将适用于许多不同的领域。这个项目属于EPSRC的统计和应用概率研究领域。这个项目由David van Dyk教授(伦敦帝国理工学院)指导，哈佛-史密森天体物理中心的一些天体物理学家也参与其中。

The aim of this thesis is to review, apply and extend current methods in Bayesian statistical modelling for astronomical data. In particular, the primary focus of this research is the modelling of the time variability of high-energy astrophysical sources such as quasars or X-ray binaries. Under this theme, the structure of the project will integrate two main research sub-components. The first sub-component will attempt to solve the problem of overlapping astronomical sources, and the method currently under development has potential applications in statistical satellite imagery processing. The astrophysics community is worried that the SpaceX satellite constellation will pollute the imagery produced by the much anticipated Legacy Survey of Space and Time (LSST) of the Vera C. Rubin Observatory. This method could be used to pre-process satellite trails off the LSST imagery and provide clean images for follow-up analysis. In addition to improving and extending the models currently used to solve this problem, the project also focuses on designing highly efficient and scalable algorithms and developing a powerful tool that astrophysicists can integrate in their current research processes. The second sub-component will attempt to solve the problem of the inconsistent Hubble constant measurements by providing a probabilistic estimation of this physical quantity of interest. This involves developing flexible and scalable methods to model astronomical time series. The novelty of the research methodology resides in the design of state-of-the-art modelling and computational techniques to draw new insights from complex data. The immediate impact of this research will be on statistical analysis of astronomical data, but the developed methods are conceived to be highly generalizable and will be applicable to many different fields. This project falls within the EPSRC "Statistics and Applied Probability research area. This project is supervised by Prof. David van Dyk (Imperial College London), and a number of astrophysicists from the Harvard-Smithsonian Center for Astrophysics are also involved.