Nuisance hardened data compression for fast likelihood-free inference

Abstract

In this paper we show how nuisance parameter marginalized posteriors can be inferred directly from simulations in a likelihood-free setting, without having to jointly infer the higher-dimensional interesting and nuisance parameter posterior first and marginalize a posteriori. The result is that for an inference task with a given number of interesting parameters, the number of simulations required to perform likelihood-free inference can be kept (roughly) the same irrespective of the number of additional nuisances to be marginalized over. To achieve this we introduce two extensions to the standard likelihood-free inference set-up. Firstly we show how nuisance parameters can be re-cast as latent variables and hence automatically marginalized over in the likelihood-free framework. Secondly, we derive an asymptotically optimal compression from $N$ data down to $n$ summaries -- one per interesting parameter -- such that the Fisher information is (asymptotically) preserved, but the summaries are insensitive (to leading order) to the nuisance parameters. This means that the nuisance marginalized inference task involves learning $n$ interesting parameters from $n$ "nuisance hardened" data summaries, regardless of the presence or number of additional nuisance parameters to be marginalized over. We validate our approach on two examples from cosmology: supernovae and weak lensing data analyses with nuisance parameterized systematics. For the supernova problem, high-fidelity posterior inference of ${\Omega }_m$ and $w_0$ (marginalized over systematics) can be obtained from just a few hundred data simulations. For the weak lensing problem, six cosmological parameters can be inferred from $\mathcal{O}(10^3)$ simulations, irrespective of whether ten additional nuisance parameters are included in the problem or not.

Submitted to MNRAS Mar 2019

References in corpus (53)

• Adam: A Method for Stochastic Optimization
• Improved cosmological constraints from a joint analysis of the SDSS-II and SNLS supernova samples
• Karhunen-Loeve eigenvalue problems in cosmology: how should we tackle large data sets?
• Power Spectrum Tomography with Weak Lensing
• Weak lensing for precision cosmology
• Cosmological parameters from lensing power spectrum and bispectrum tomography
• Effects of Baryons and Dissipation on the Matter Power Spectrum
• Weak Lensing and Cosmology
• Galaxy alignments: An overview
• Massive Lossless Data Compression and Multiple Parameter Estimation from Galaxy Spectra
• Dark Synergy: Gravitational Lensing and the CMB
• Fast likelihood-free cosmology with neural density estimators and active learning
• Origins of weak lensing systematics, and requirements on future instrumentation (or knowledge of instrumentation)
• Constraining the thick disc formation scenario of the Milky Way
• Mining gold from implicit models to improve likelihood-free inference
• Constraining Effective Field Theories with Machine Learning
• Massive optimal data compression and density estimation for scalable, likelihood-free inference in cosmology
• A Guide to Constraining Effective Field Theories with Machine Learning
• Probing Cosmology with Weak Lensing Peak Counts
• Non-Gaussian information from weak lensing data via deep learning
• Weak lensing power spectra for precision cosmology: Multiple-deflection, reduced shear and lensing bias corrections
• Baryons, Neutrinos, Feedback and Weak Gravitational Lensing
• Approximate Bayesian Computation for Astronomical Model Analysis: A Case Study in Galaxy Demographics and Morphological Transformation at High Redshift
• Automatic physical inference with information maximising neural networks
• Likelihood-Free Cosmological Inference with Type Ia Supernovae: Approximate Bayesian Computation for a Complete Treatment of Uncertainty
• Generalized massive optimal data compression
• Towards emulating cosmic shear data: Revisiting the calibration of the shear measurements for the Kilo-Degree Survey
• Approximate Bayesian Computation for Forward Modeling in Cosmology
• The Limits of Cosmic Shear
• CARS: The CFHTLS-Archive-Research Survey III. First detection of cosmic magnification in samples of normal high-z galaxies
• cosmoabc: Likelihood-free inference via Population Monte Carlo Approximate Bayesian Computation
• Cosmological constraints from noisy convergence maps through deep learning
• Bayesian optimisation for likelihood-free cosmological inference
• Weak Gravitational Lensing Bispectrum
• Likelihood-free inference with emulator networks
• A new model to predict weak-lensing peak counts II. Parameter constraint strategies
• On the complementarity of galaxy clustering with cosmic shear and flux magnification
• Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows
• Inferring the mass of sub-millimetre galaxies by exploiting their gravitational magnification of background galaxies
• A New Method to Measure the Post-Reionization Ionizing Background from the Joint Distribution of Lyman-$\alpha$ and Lyman-$\beta$ Forest Transmission
• Shear and Magnification: Cosmic Complementarity
• Approximate Bayesian Computation in Large Scale Structure: constraining the galaxy-halo connection
• The skewed weak lensing likelihood: why biases arise, despite data and theory being sound
• Weak Lensing with Sizes, Magnitudes and Shapes
• The redshift distribution of cosmological samples: a forward modeling approach
• Combining Size and Shape in Weak Lensing
• Weak lensing peak statistics in the era of large scale cosmological surveys
• Accelerating Approximate Bayesian Computation with Quantile Regression: Application to Cosmological Redshift Distributions
• Cosmological model discrimination with Deep Learning
• Inferring the photometric and size evolution of galaxies from image simulations
• Extreme data compression for the CMB
• Optimal proposals for Approximate Bayesian Computation
• The many flavours of photometric redshifts

Cited by in corpus (78)

• Machine learning and the physical sciences
• The frontier of simulation-based inference
• The Quijote simulations
• The CAMELS project: Cosmology and Astrophysics with MachinE Learning Simulations
• Fast likelihood-free cosmology with neural density estimators and active learning
• Mining gold from implicit models to improve likelihood-free inference
• Cosmological constraints with deep learning from KiDS-450 weak lensing maps
• Likelihood-free inference with neural compression of DES SV weak lensing map statistics
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• Mining for Dark Matter Substructure: Inferring subhalo population properties from strong lenses with machine learning
• SPECULATOR: Emulating stellar population synthesis for fast and accurate galaxy spectra and photometry
• Super-resolution emulator of cosmological simulations using deep physical models
• Simulation-Based Inference of Reionization Parameters From 3D Tomographic 21 cm Lightcone Images
• Uncertainty Aware Learning for High Energy Physics
• Fast and Credible Likelihood-Free Cosmology with Truncated Marginal Neural Ratio Estimation
• Translation and Rotation Equivariant Normalizing Flow (TRENF) for Optimal Cosmological Analysis
• Implicit Likelihood Inference of Reionization Parameters from the 21 cm Power Spectrum
• Kernel-Based Emulator for the 3D Matter Power Spectrum from CLASS
• The sum of the masses of the Milky Way and M31: a likelihood-free inference approach
• A Cautionary Tale of Decorrelating Theory Uncertainties
• Euclid: The reduced shear approximation and magnification bias for Stage IV cosmic shear experiments
• Forward modeling of galaxy populations for cosmological redshift distribution inference
• Cosmology with cosmic web environments I. Real-space power spectra
• Simulation-based inference of dynamical galaxy cluster masses with 3D convolutional neural networks
• Likelihood methods for CMB experiments
• Measurement of the B-band Galaxy Luminosity Function with Approximate Bayesian Computation
• Likelihood-free Cosmological Constraints with Artificial Neural Networks: An Application on Hubble Parameters and SNe Ia
• KiDS-SBI: Simulation-based inference analysis of KiDS-1000 cosmic shear
• Robust Simulation-Based Inference in Cosmology with Bayesian Neural Networks
• EFTofLSS meets simulation-based inference: ${\sigma }_8$ from biased tracers
• Quijote-PNG: Quasi-maximum likelihood estimation of Primordial Non-Gaussianity in the non-linear dark matter density field
• SLITronomy: towards a fully wavelet-based strong lensing inversion technique
• Non-Gaussian modelling and statistical denoising of Planck dust polarization full-sky maps using scattering transforms
• MillimeterDL: Deep Learning Simulations of the Microwave Sky
• The PAU Survey: Measurement of Narrow-band galaxy properties with Approximate Bayesian Computation
• Fisher Forecasts for Primordial non-Gaussianity from Persistent Homology
• Primordial non-Gaussianity and non-Gaussian Covariance
• 21st Century Statistical and Computational Challenges in Astrophysics
• pop-cosmos: A comprehensive picture of the galaxy population from COSMOS data
• Calibrating cosmological simulations with implicit likelihood inference using galaxy growth observables
• Unbiased likelihood-free inference of the Hubble constant from light standard sirens
• CoLFI: Cosmological Likelihood-free Inference with Neural Density Estimators
• A simulation-based inference pipeline for cosmic shear with the Kilo-Degree Survey
• Bayesian Parameterization of Continuum Battery Models from Featurized Electrochemical Measurements Considering Noise
• Data Compression and Inference in Cosmology with Self-Supervised Machine Learning
• Cosmology with Persistent Homology: a Fisher Forecast
• Scalable inference with Autoregressive Neural Ratio Estimation
• Quantification of high dimensional non-Gaussianities and its implication to Fisher analysis in cosmology
• Machine Learning and Cosmology
• Cosmological Parameter Estimation and Inference using Deep Summaries
• Reconstructing Probability Distributions with Gaussian Processes
• Flow-Based Likelihoods for Non-Gaussian Inference
• Likelihood-free Forward Modeling for Cluster Weak Lensing and Cosmology
• Exact joint likelihood of pseudo-$C_{\ell }$ estimates from correlated Gaussian cosmological fields
• Estimation of HII Bubble Size Distribution from 21cm Power Spectrum with Artificial Neural Networks
• Bias and Priors in Machine Learning Calibrations for High Energy Physics
• Bayesian evidence-driven likelihood selection for sky-averaged 21-cm signal extraction
• SIDE-real: Supernova Ia Dust Extinction with truncated marginal neural ratio estimation applied to real data
• Catalog-free modeling of galaxy types in deep images: Massive dimensional reduction with neural networks
• Cosmological constraints from low redshift 21 cm intensity mapping with machine learning
• Full-waveform earthquake source inversion using simulation-based inference
• Dealing with Nuisance Parameters using Machine Learning in High Energy Physics: a Review
• Simulation-based Anomaly Detection for Multileptons at the LHC
• A Simulation Based Inference Approach to the Dynamics of the MW-LMC System -- Validation
• COmoving Computer Acceleration (COCA): $N$-body simulations in an emulated frame of reference
• Inferring Cosmological Parameters on SDSS via Domain-Generalized Neural Networks and Lightcone Simulations
• Simulation-based inference has its own Dodelson-Schneider effect (but it knows that it does)
• Comparison of Bayesian inference methods using the Loreli II database of hydro-radiative simulations of the 21-cm signal
• The RSD Sorting Hat: Unmixing Radial Scales in Projection
• Radio pulsar population synthesis with consistent flux measurements using simulation-based inference
• Galaxy mass profiles with convolutional neural networks
• Impact of Redshift Space Distortion on Persistent Homology of cosmic matter density field
• STAR NRE: Solving supernova selection effects with set-based truncated auto-regressive neural ratio estimation
• Optimal data compression for Lyman-$\alpha$ forest cosmology
• Diagnosing Systematic Effects Using the Inferred Initial Power Spectrum
• Graph-based Summary Statistics for Revealing the Stochastic Gravitational Wave Background in Pulsar Timing Arrays
• $\text{SBi3PCF:}$ Simulation-based inference with the integrated 3PCF
• GalSBI-SPS: a stellar population synthesis-based galaxy population model for cosmology and galaxy evolution applications