MultiNest: an efficient and robust Bayesian inference tool for cosmology and particle physics

By Farhan Feroz, M. Hobson, M. Bridges. 2008

2,499 citations recorded by the source corpus.

We present further development and the first public release o f our multimodal nested sampling algorithm, called MULTINEST. This Bayesian inference tool calculates the evidence, with an associated error estimate, and produces posterior s amples from distributions that may contain multiple modes and pronounced (curving) degeneracies in high dimensions. The developments presented here lead to further substantia l improvements in sampling efficiency and robustness, as compared to the original algorit hm presented in Feroz & Hobson (2008), which itself significantly outperformed existi ng MCMC techniques in a wide range of astrophysical inference problems. The accuracy and economy of the MULTINEST algorithm is demonstrated by application to two toy problems and to a cosmological in

Subfields: Artificial Intelligence, Nuclear and High Energy Physics, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Nuclear and High Energy Physics

In Statistics and Probability

Citations in the corpus

STELLAR AND PLANETARY PROPERTIES OF K2 CAMPAIGN 1 CANDIDATES AND VALIDATION OF 17 PLANETS, INCLUDING A PLANET RECEIVING EARTH-LIKE INSOLATION (2015, methodology): “MULTINEST is designed to sample multimodal posteriors, where other samplers such as MCMC algorithms often struggle.”
Stellar Population Inference with Prospector (2020, methodology): “Nested sampling in Prospector uses the dynesty pure-Python code package (Speagle 2020), an implementation of algorithms described in Feroz et al. (2009); Higson et al. (2019).”
THE UV LUMINOSITY FUNCTION OF STAR-FORMING GALAXIES VIA DROPOUT SELECTION AT REDSHIFTS z ∼ 7 AND 8 FROM THE 2012 ULTRA DEEP FIELD CAMPAIGN (2012, methodology): “This calculation amounts to an Nbinparameter estimation problem, and we use the MultiNest nested sampling code for Bayesian inference problems (Feroz & Hobson 2008; Feroz et al. 2009) to maximize the Φk likelihoods.”
KiDS+VIKING-450: Cosmic shear tomography with optical and infrared data (2018, methodology): “This setup is based on the nested sampling algorithm MultiNest (Feroz et al. 2009) as implemented in the python wrapper PyMultiNest (Buchner et al. 2014) that is included in the MontePython package (Audren et al. 2013; Brinckmann & Lesgourgues 2018).”
THE INFLUENCE OF NONUNIFORM CLOUD COVER ON TRANSIT TRANSMISSION SPECTRA (2015, methodology): “We employ PyMultiNest4 (Buchner et al. 2014), a python wrapper to the commonly used MultiNest (Feroz et al. 2009) nested sampling algorithm, to explore the parameter degeneracies.”
The NANOGrav 11-year Data Set: High-precision Timing of 45 Millisecond Pulsars (2017, methodology): “This noise model is incorporated into a joint likelihood containing all timing model parameters, and run through an MCMC inference package44 and through MultiNest (Feroz et al. 2009) to produce maximum-likelihood parameter estimates and Bayesian evidence for the presence of red noise, respectively.”
THE NANOGRAV NINE-YEAR DATA SET: OBSERVATIONS, ARRIVAL TIME MEASUREMENTS, AND ANALYSIS OF 37 MILLISECOND PULSARS (2015, methodology): “The red noise model selection is performed with MultiNest (Feroz et al. 2009) using a Bayes factor threshold of 100 to determine whether red noise is included in the final model.”
Curvature tension: Evidence for a closed universe (2019, background): “[26] F.”
The connection between stellar granulation and oscillation as seen by the Kepler mission (2014, methodology): “To fit the models to the power density spectra we have used a Bayesian inference tool (MultiNest; Feroz et al. 2009) assuming that the difference between the model and observed spectrum follow a χ2statistic with two degrees of freedom (Gabriel 1994)5.”
Exploring cosmic origins with CORE: Inflation (2016, methodology): “All evidences have then been derived using the MULTINEST nested sampling algorithm [180, 181, 239] with a target accuracy of 10−4 and the number of live points equal to 30000.”
Two-dimensional kinematics of SLACS lenses – III. Mass structure and dynamics of early-type lens galaxies beyond z ≃ 0.1 (2011, methodology): “…the posterior distribution Pr(η | d) is accomplished by making use of the very efficient and robust MultiNest algorithm (Feroz & Hobson 2008; Feroz et al. 2009), which implements the ‘nested sampling’ Monte Carlo technique (Skilling 2004; Sivia & Skilling 2006), and can provide reliable…”
X-ray spectral modelling of the AGN obscuring region in the CDFS: Bayesian model selection and catalogue (2014, background): “MultiNest (Feroz et al. 2009) solves this problem efficiently by clustering the live points into multidimensional ellipses, and drawing from these subspaces under the assumption that higher-valued points can only be found in proximity of already drawn points.”
DWARF GALAXY ANNIHILATION AND DECAY EMISSION PROFILES FOR DARK MATTER EXPERIMENTS (2014, methodology): “…the parameter space efficiently, we use the nested-sampling Monte Carlo algorithm introduced by Skilling (2004) and implemented in the software package MultiNest (Feroz & Hobson 2008; Feroz et al. 2009), which outputs samples from the model’s posterior probability distribution function (PDF).”