Unveiling Cosmic Dawn and the Epoch of Reionisation with HERA

该项目的目的是促进发展和应用先进的统计方法来分析氢再电离阵列（HERA）现有和未来的数据。关键问题包括：在明亮前景辐射存在的情况下，高效、精确地估计21cm波动的功率谱。校正由于消除无线电频率干扰而导致的数据减少所造成的“泄漏”。通过严格应用零检验并通过分析管道提供模拟数据，分析结果对各种分析假设的稳定性。准确估计数据的协方差矩阵，优化提取微弱的21cm信号。以计算效率高的方式建模/模拟物理EoR和宇宙黎明信号。其中一些工作将与HERA协作的其他成员合作进行，这些成员分别位于伯克利、麦吉尔、西开普省、喷气推进实验室、宾夕法尼亚大学和华盛顿大学。学生将学习分析和解释射电天文数据的技能，将高维推理技术应用于大量数据，将结构化和无监督机器学习技术应用于检测数据中的虚假特征，以及开发可在高性能计算机上运行的快速并行数据分析代码。

The aim of this project is to contribute to the development and application of advanced statistical methods to the analysis of existing and future data from the Hydrogen Epoch of Reionization Array (HERA). Key problems include: Efficiently and precisely estimating the power spectrum of 21cm fluctuations, in the presence of bright foreground emission. Correcting for "leakage" due to cuts in the data, imposed by excising radio frequency interference. Analysing the stability of the results to various analysis assumptions, by rigorously applying null tests and feeding simulated data through the analysis pipeline. Accurately estimating the covariance matrix of the data, which is needed to optimally extract the faint 21cm signal. Modelling/simulating the physical EoR and Cosmic Dawn signal in a computationally-efficient way.Some of this work will be carried out in collaboration with other members of the HERA Collaboration based in Berkeley, McGill, U. Western Cape, JPL, U. Penn, and U. Washington. The student will gain skills in analysis and interpretation of radio astronomical data, the application of high-dimensional inference techniques to large volumes of data, the application of structured and unsupervised machine learning techniques to detect spurious features in data, and the development of fast parallel data analysis codes to be run on high-performance computers.