Understanding a Data-Rich Universe with Data-Driven Approaches

• Activity Number: 163 - Understanding a Data-Rich Universe with Data-Driven Approaches
• Type: Topic-Contributed
• Date/Time: Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Boris Leistedt* and Aarya Patil* and Stephen Portillo* and Miles Cranmer* and Kaisey Mandel*
• Companies: Imperial College London and University of Toronto and University of Washington and Princeton University and University of Cambridge
• Keywords: astronomy; astrostatistics; machine learning; data-driven; hierarchical models

Abstract

Astrophysics is currently facing a deluge of data of increasing scope and complexity. As theoretical models and traditional methods struggle to keep pace, there has been substantial interest in using flexible, data-driven methods at scale to infer the properties of underlying astronomical populations. This session will highlight applications of data-driven approaches in recent years to understanding stars and galaxies across a wide variety of datasets, with a focus on: (1) how unique problems in astronomy have led to the development and application of novel methods from machine learning and elsewhere, (2) how these methods are integrated into larger statistical models to account for observational uncertainties, missing data, and physical constraints, and (3) how the results can be used to better understand underlying astrophysical populations.