Uncertainty Quantification in Astronomy

Aneta Siemiginowska Chair
Harvard-Smithsonian Center for Astrophysics
David van Dyk Discussant
Imperial College London
Aneta Siemiginowska Organizer
Harvard-Smithsonian Center for Astrophysics
Gwendolyn Eadie Organizer
University of Toronto
Monday, Aug 7: 2:00 PM - 3:50 PM
Topic-Contributed Paper Session 
Metro Toronto Convention Centre 
Room: CC-206B 



Main Sponsor

International Statistical Institute

Co Sponsors

Astrostatistics Interest Group
Section on Physical and Engineering Sciences


Accounting for uncertainty in complex simulation-based models and emerging big data methodologies


Kyle Cranmer, University of Wisconsin-Madison

On computationally efficient methods for testing multivariate distributions with unknown parameters

Despite the popularity of classical goodness fit tests such as Pearson's chi-squared and Kolmogorov-Smirnov, their applicability often faces serious challenges in practical applications. For instance, in a binned data regime, low counts may affect the validity of the asymptotic results. Excessively large bins, on the other hand, may lead to loss of power. In the unbinned data regime, tests such as Kolmogorov-Smirnov and Cramer-von Mises do not enjoy distribution-freeness if the models under study are multivariate and/or involve unknown parameters. As a result, one needs to simulate the distribution of the test statistic on a case-by-case basis. In this talk, I will discuss a testing strategy that allows us to overcome these shortcomings and equips experimentalists with a novel tool to perform goodness-of-fit while reducing substantially the computational costs. 


Sara Algeri, University of Minnesota

Transforming the first eROSITA X-ray all-sky survey into astrophysics knowledge

Large astronomical surveys of millions of sources such as eROSITA's
X-ray all-sky survey attempt to constrain the physical mechanisms at
play in distant galaxies. This talk will give an overview of some of the
statistical challenges in astrophysics, in particular uncertainty
estimation within and deciding between competing, physical, parametric
models. I will discuss which inference problems can be addressed with
machine learning and when astronomy is driven to custom Bayesian models
analysed with nested sampling or Hamiltonian Monte Carlo. Two highlights
from X-ray analysis include detection of significant variability in
Poisson data, and discovering accretion onto supermassive black holes
veiled in thick gas and dust. 


Johannes Buchner, Max Planck Institute for Extraterrestrial Physics