We present a new, fully generative model of optical telescope image sets, along with a variational procedure for inference. Each pixel intensity is treated as a Poisson random variable, with a rate parameter dependent on latent properties of stars and galaxies. Key latent properties are themselves random, with scientific prior distributions constructed from large ancillary data sets. We check our approach on synthetic images. We also run it on images from a major sky survey, where it exceeds the performance of the current state-of-the-art method for locating celestial bodies and measuring their colors.
@conference{regier2015celeste, year = {2015}, author = {Regier, Jeffrey and Miller, Andrew C. and McAuliffe, Jon and Adams, Ryan P. and Hoffman, Matthew D. and Lang, Dustin and Schlegel, David and Prabhat}, title = {Celeste: Variational Inference for a Generative Model of Astronomical Images}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning (ICML)}, note = {arXiv:1506.01351 [astro-ph.IM]}, keywords = {ICML, astronomy, variational inference, graphical models} }