Fast and Credible Likelihood-Free Cosmology with Truncated Marginal Neural Ratio Estimation

Abstract

Sampling-based inference techniques are central to modern cosmological data analysis; these methods, however, scale poorly with dimensionality and typically require approximate or intractable likelihoods. In this paper we describe how Truncated Marginal Neural Ratio Estimation (TMNRE) (a new approach in so-called simulation-based inference) naturally evades these issues, improving the $(i)$ efficiency, $(ii)$ scalability, and $(iii)$ trustworthiness of the inferred posteriors. Using measurements of the Cosmic Microwave Background (CMB), we show that TMNRE can achieve converged posteriors using orders of magnitude fewer simulator calls than conventional Markov Chain Monte Carlo (MCMC) methods. Remarkably, the required number of samples is effectively independent of the number of nuisance parameters. In addition, a property called \emph{local amortization} allows the performance of rigorous statistical consistency checks that are not accessible to sampling-based methods. TMNRE promises to become a powerful tool for cosmological data analysis, particularly in the context of extended cosmologies, where the timescale required for conventional sampling-based inference methods to converge can greatly exceed that of simple cosmological models such as $\Lambda$CDM. To perform these computations, we use an implementation of TMNRE via the open-source code \texttt{swyft}.

v2: accepted journal version. v1: 37 pages, 13 figures. \texttt{swyft} is available at https://github.com/undark-lab/swyft, and demonstration code for cosmological examples is available at https://github.com/acole1221/swyft-CMB

References in corpus (26)

• Array Programming with NumPy
• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• The Clustering of the SDSS DR7 Main Galaxy Sample I: A 4 per cent Distance Measure at z=0.15
• Massive neutrinos and cosmology
• Probing cosmological parameters with the CMB: Forecasts from full Monte Carlo simulations
• Neutrino mass from Cosmology
• Fast likelihood-free cosmology with neural density estimators and active learning
• Neutrino cosmology and Planck
• Nuisance hardened data compression for fast likelihood-free inference
• Flexible statistical inference for mechanistic models of neural dynamics
• Automatic Posterior Transformation for Likelihood-Free Inference
• astroABC: An Approximate Bayesian Computation Sequential Monte Carlo sampler for cosmological parameter estimation
• Likelihood methods for the combined analysis of CMB temperature and polarisation power spectra
• sCOLA: The N-body COLA Method Extended to the Spatial Domain
• NECOLA: Towards a Universal Field-level Cosmological Emulator
• Parameter Inference for Weak Lensing using Gaussian Processes and MOPED
• Single frequency CMB B-mode inference with realistic foregrounds from a single training image
• Multifield Cosmology with Artificial Intelligence
• Extreme data compression while searching for new physics
• Solving high-dimensional parameter inference: marginal posterior densities & Moment Networks
• Targeted Likelihood-Free Inference of Dark Matter Substructure in Strongly-Lensed Galaxies
• Simulation-efficient marginal posterior estimation with swyft: stop wasting your precious time
• Arbitrary Marginal Neural Ratio Estimation for Simulation-based Inference
• Pico: Parameters for the Impatient Cosmologist
• A Python compressed low-$\ell$ Planck likelihood for temperature and polarization
• Robust marginalization of baryonic effects for cosmological inference at the field level

Cited by in corpus (32)

• The CAMELS project: public data release
• Neural posterior estimation for exoplanetary atmospheric retrieval
• Robust Field-level Likelihood-free Inference with Galaxies
• Exploring the likelihood of the 21-cm power spectrum with simulation-based inference
• LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology
• 21cmEMU: an emulator of 21cmFAST summary observables
• Constraining the X-ray heating and reionization using 21-cm power spectra with Marginal Neural Ratio Estimation
• EFTofLSS meets simulation-based inference: ${\sigma }_8$ from biased tracers
• Primordial non-Gaussianity and non-Gaussian Covariance
• Fisher Forecasts for Primordial non-Gaussianity from Persistent Homology
• Bayesian model comparison for simulation-based inference
• One never walks alone: the effect of the perturber population on subhalo measurements in strong gravitational lenses
• Cosmology with Persistent Homology: a Fisher Forecast
• Scalable inference with Autoregressive Neural Ratio Estimation
• Fast Likelihood-free Reconstruction of Gravitational Wave Backgrounds
• Simulation-Based Inference of the sky-averaged 21-cm signal from CD-EoR with REACH
• A comparison of Bayesian sampling algorithms for high-dimensional particle physics and cosmology applications
• Fast likelihood-free inference in the LSS Stage IV era
• Inferring subhalo effective density slopes from strong lensing observations with neural likelihood-ratio estimation
• Enhancing Cosmological Model Selection with Interpretable Machine Learning
• MF-Box: Multi-fidelity and multi-scale emulation for the matter power spectrum
• Near-instantaneous Atmospheric Retrievals and Model Comparison with FASTER
• Simulation-based inference has its own Dodelson-Schneider effect (but it knows that it does)
• Reconstructing axion-like particles from beam dumps with simulation-based inference
• Bayesian technique to combine independently-trained Machine-Learning models applied to direct dark matter detection
• STAR NRE: Solving supernova selection effects with set-based truncated auto-regressive neural ratio estimation
• Towards characterizing dark matter subhalo perturbations in stellar streams with graph neural networks
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• Finding excesses in model parameter space
• How to embed any likelihood into SBI: Application to Planck + Stage IV galaxy surveys and Dynamical Dark Energy
• Calibrating Bayesian Tension Statistics using Neural Ratio Estimation
• Fast Sampling of Cosmological Initial Conditions with Gaussian Neural Posterior Estimation