CAREER: Building the Merger Tree of the Milky Way with Machine Learning

首席研究员（PI） Necib和她的团队将在我们的星系中寻找恒星碎片。这些恒星结构是星系与银河系相撞时形成的。该团队将利用人工智能开发一个自动化过程，以完成分析来自欧洲航天局（ESA）盖亚任务的超过15亿颗恒星的观测结果的艰巨任务。该团队将把他们的数据与模拟结果进行比较，绘制出银河系形成的历史。PI致力于扩大参与。在她成功的突尼斯播客Ga3da Falakia的基础上，PI将与她的研究生合作，用多种语言制作一系列播客。PI还将为本科生和高中生举办一系列盖亚黑客马拉松。PI的目标是使用监督和无监督机器学习技术构建银河系的合并树。PI将构建一个集成测量不确定性和随数据量增加而增加的尺度的聚类算法，以及一个可解释的等变图神经网络，预测合并时间和银河系的合并质量，在Illustris-TNG50上训练，并应用于Gaia DR3/4数据。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Principal Investigator (PI) Necib and her team will search for stellar debris in our galaxy. These stellar structures were formed when galaxies collided with the Milky Way. The team will use artificial intelligence to develop an automated process to complete the difficult task of analyzing observations of more than 1.5 billion stars from the European Space Agency (ESA) mission Gaia. The team will compare their data to simulations to chart the history of the formation of the Milky Way. The PI is committed to broadening participation. Building on her successful Tunisian Podcast Ga3da Falakia, the PI will work with her graduate students to produce a series of podcasts in a range of languages. The PI will also host a series of Gaia Hackathons for undergraduate and high school students. The PI aims to build the merger tree of the Milky Way using both supervised and unsupervised machine learning techniques. The PI will build a clustering algorithm that integrates measurement uncertainties and scales with the increase in data size, as well as an interpretable equivariant graph neural network that predicts the merging times and infall masses of mergers into the Milky Way, trained on Illustris-TNG50, and applied to Gaia DR3/4 data.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.