CosmoSIS: modular cosmological parameter estimation

TL;DR

CosmoSIS introduces a modular framework for cosmological parameter estimation that connects diverse, multi-language code via a shared DataBlock to form flexible pipelines. It details the architecture, data flow, and a standard library, and demonstrates practical pipelines including CAMB, Planck, CFHTLenS, BICEP2, and JLA analyses. The work discusses the benefits of modularity—easy model substitution, verifiability, cross-probe consistency, and collaborative sharing—along with the associated costs and implementation considerations, and it promotes community-driven module repositories to foster reproducible, scalable cosmological inference.

Abstract

Cosmological parameter estimation is entering a new era. Large collaborations need to coordinate high-stakes analyses using multiple methods; furthermore such analyses have grown in complexity due to sophisticated models of cosmology and systematic uncertainties. In this paper we argue that modularity is the key to addressing these challenges: calculations should be broken up into interchangeable modular units with inputs and outputs clearly defined. We present a new framework for cosmological parameter estimation, CosmoSIS, designed to connect together, share, and advance development of inference tools across the community. We describe the modules already available in CosmoSIS, including CAMB, Planck, cosmic shear calculations, and a suite of samplers. We illustrate it using demonstration code that you can run out-of-the-box with the installer available at http://bitbucket.org/joezuntz/cosmosis

Figures (8)

• Figure 1: A schematic of the modular calculations for a galaxy power spectrum likelihood.
• Figure 2: CosmoSIS demo one output plot, showing CMB spectra output from camb. CMB spectra and a host of other cosmology theory values are saved from the cambCosmoSIS module for later pipeline modules to use.
• Figure 3: CosmoSIS demo 3 output plot, showing the constraints on the primordial tensor fraction $r$ from BICEP2 B-mode data. The y-axis is the normalized likelihood and the vertical lines show 68% and 95% contours.
• Figure 4: An example CosmoSIS constraint on the JLA supernova data set - all the 1D and 2D constraint plots are generated by CosmoSIS; this example ( CosmoSIS demo 5) shows constraints from the JLA SDSS supernova sample on the Hubble parameter $h$ and the supernova magnitude parameter $\Delta M$ made using the emcee sampler.
• Figure 5: The CosmoSIS test sampler produces and saves all the cosmological outputs for a set of parameters, and they can immediately be plotted with the postprocess program. This example from CosmoSIS demo six shows the cosmic shear spectra generated for the CFHTLenS redshift bins.