Unsupervised Machine Learning in Astrophysical Data

Abstract

Large astrophysical surveys provide us with a wealth of data that has important implications for fundamental physics questions. Fully exploiting these datasets requires new techniques and approaches, including machine learning. However, it is difficult to completely simulate many astrophysical systems. As a result, unsupervised learning techniques can be particularly useful for these astrophysical questions. I discuss two applications of unsupervised machine learning to astrophysical data: discovering stellar streams and measuring the local Galactic density.