Computational statistics using the Bayesian Inference Engine

This paper introduces the Bayesian Inference Engine (BIE), a general parallel-optimised software package for parameter inference and model selection. This package is motivated by the analysis needs of modern astronomical surveys and the need to organise and reuse expensive derived data. I describe key concepts that illustrate the power of Bayesian inference to address these needs and outline the computational challenge. The techniques presented are based on experience gained in modelling star-counts and stellar populations, analysing the morphology of galaxy images, and performing Bayesian investigations of semi-analytic models of galaxy formation. These inference problems require advanced Markov chain Monte Carlo (MCMC) algorithms that expedite sampling, mixing, and the analysis of the Bayesian posterior distribution. The BIE was designed to be a collaborative platform for applying Bayesian methodology to astronomy. By providing a variety of statistical algorithms for all phases of the inference problem, a user may explore a variety of approaches with a single model implementation. Indeed, each of the separate scientific investigations above has benefited from the solutions posed for the other investigations, and I anticipate that the same solutions will be of general value for other areas of astronomical research. Finally, to protect one’s computational investment against loss any equipment failure and human error, the BIE includes a comprehensive persistence system that enables byte-level checkpointing and restoration. Additional technical details and download details are available from http://www.astro.umass.edu/bie. The BIE is distributed under the GNU GPL.