Probing Galaxy Evolution Through Time with Deep Learning

整个宇宙中星系的形状和动力学提供了一个强大的透镜，通过它我们可以了解塑造它们的过程。准确地标记星系的形态是星系演化研究中的一项重要任务，直到最近，这项任务还最好由人类来完成。公民科学项目，如银河动物园，为天文学家提供了数十万张有标签的星系图像，揭示了其中复杂的结构和历史。但是，在即将到来的大数据天文学时代，机器学习和人工智能正在成为处理大规模下一代天空测量的流行替代方案。该博士项目将致力于进一步开发先进的图像识别和目标检测技术。它与尖端的深度成像技术合作，如遗产空间和时间调查，以及超至尊凸轮上的斯巴鲁战略计划，它将产生数百万星系的机器标记形态目录，并揭示宇宙时间中星系结构的变化性质。

The shapes and dynamics of galaxies throughout the Universe provide a powerful lens through which we can understand the processes that shape them. Accurately labelling the morphologies of galaxies is a vital task in the study of galaxy evolution, and is a task that until recently was best completed by humans. Citizen science projects, such as Galaxy Zoo, have provided astronomers with hundreds of thousands of labelled galaxy images, revealing the intricate structures and histories held within. But, in the coming era of Big Data Astronomy, machine learning and AI are becoming a popular alternative to handle the vast scale of next-generation sky surveys. This PhD project will aim to further develop advanced image recognition and object detection technologies. Working with cutting-edge deep imaging, such as the Legacy Survey of Space and Time, and the Subaru Strategic Program on the Hyper-Supreme Cam, it will produce machine-labelled morphological catalogues of millions of galaxies, and uncover the changing nature of galaxy structure through cosmic time.