The frontier of simulation-based inference

Abstract

Many domains of science have developed complex simulations to describe phenomena of interest. While these simulations provide high-fidelity models, they are poorly suited for inference and lead to challenging inverse problems. We review the rapidly developing field of simulation-based inference and identify the forces giving new momentum to the field. Finally, we describe how the frontier is expanding so that a broad audience can appreciate the profound change these developments may have on science.

arXiv comment10 pages, 3 figures, proceedings for the Sackler Colloquia at the US National Academy of Sciences. v2: fixed typos. v3: clarified text, added references

References in corpus (37)

• WaveNet: A Generative Model for Raw Audio
• Asymptotic formulae for likelihood-based tests of new physics
• Machine learning and the physical sciences
• Conditional Image Generation with PixelCNN Decoders
• NICE: Non-linear Independent Components Estimation
• Variational Inference with Normalizing Flows
• MADE: Masked Autoencoder for Distribution Estimation
• Sequential design of computer experiments for the estimation of a probability of failure
• Fast likelihood-free cosmology with neural density estimators and active learning
• Masked Autoregressive Flow for Density Estimation
• 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
• Fast $\epsilon$-free Inference of Simulation Models with Bayesian Conditional Density Estimation
• Approximating Likelihood Ratios with Calibrated Discriminative Classifiers
• QCD-Aware Recursive Neural Networks for Jet Physics
• Nuisance hardened data compression for fast likelihood-free inference
• JUNIPR: a Framework for Unsupervised Machine Learning in Particle Physics
• Generalized massive optimal data compression
• GPS-ABC: Gaussian Process Surrogate Approximate Bayesian Computation
• Hierarchical Implicit Models and Likelihood-Free Variational Inference
• Flexible statistical inference for mechanistic models of neural dynamics
• Multi-fidelity Bayesian Optimisation with Continuous Approximations
• Recurrent Inference Machines for Solving Inverse Problems
• Neural Networks for Full Phase-space Reweighting and Parameter Tuning
• Bayesian Optimization for Likelihood-Free Inference of Simulator-Based Statistical Models
• Likelihood-free inference with emulator networks
• Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows
• INFERNO: Inference-Aware Neural Optimisation
• Inference Networks for Sequential Monte Carlo in Graphical Models
• ABC-CDE: Towards Approximate Bayesian Computation with Complex High-Dimensional Data and Limited Simulations
• Inference Compilation and Universal Probabilistic Programming
• Etalumis: Bringing Probabilistic Programming to Scientific Simulators at Scale
• Likelihood-free MCMC with Amortized Approximate Ratio Estimators
• Dynamic Likelihood-free Inference via Ratio Estimation (DIRE)
• Optimization Monte Carlo: Efficient and Embarrassingly Parallel Likelihood-Free Inference

Cited by in corpus (101)

• On scientific understanding with artificial intelligence
• Normalizing Flows for Probabilistic Modeling and Inference
• Real-time gravitational-wave science with neural posterior estimation
• Bayesian parameter estimation using conditional variational autoencoders for gravitational-wave astronomy
• Technology Readiness Levels for Machine Learning Systems
• Mining gold from implicit models to improve likelihood-free inference
• Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference
• Model calibration using ESEm v1.0.0 -- an open, scalable Earth System Emulator
• Machine learning for weather and climate are worlds apart
• Fast and Credible Likelihood-Free Cosmology with Truncated Marginal Neural Ratio Estimation
• Publishing statistical models: Getting the most out of particle physics experiments
• Translation and Rotation Equivariant Normalizing Flow (TRENF) for Optimal Cosmological Analysis
• FlowPM: Distributed TensorFlow Implementation of the FastPM Cosmological N-body Solver
• Recovery of TESS Stellar Rotation Periods Using Deep Learning
• Field Level Neural Network Emulator for Cosmological N-body Simulations
• Learning physical properties of anomalous random walks using graph neural networks
• A neural simulation-based inference approach for characterizing the Galactic Center $\gamma$-ray excess
• Lossless, Scalable Implicit Likelihood Inference for Cosmological Fields
• Learning to Simulate High Energy Particle Collisions from Unlabeled Data
• Accurate Machine Learning Atmospheric Retrieval via a Neural Network Surrogate Model for Radiative Transfer
• Normalizing flows as an avenue to study overlapping gravitational wave signals
• Accelerated Bayesian SED Modeling using Amortized Neural Posterior Estimation
• Implicit Likelihood Inference of Reionization Parameters from the 21 cm Power Spectrum
• OutbreakFlow: Model-based Bayesian inference of disease outbreak dynamics with invertible neural networks and its application to the COVID-19 pandemics in Germany
• A Ubiquitous Unifying Degeneracy in Two-Body Microlensing Systems
• SICRET: Supernova Ia Cosmology with truncated marginal neural Ratio EsTimation
• $\mathrm{S}\mathrm{IM}\mathrm{BIG}$: Mock Challenge for a Forward Modeling Approach to Galaxy Clustering
• Simulation-based inference of dynamical galaxy cluster masses with 3D convolutional neural networks
• Bayesian Stokes inversion with Normalizing flows
• Real-Time Likelihood-Free Inference of Roman Binary Microlensing Events with Amortized Neural Posterior Estimation
• Towards constraining warm dark matter with stellar streams through neural simulation-based inference
• Robust Simulation-Based Inference in Cosmology with Bayesian Neural Networks
• Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena
• Deducing Neutron Star Equation of State Parameters Directly From Telescope Spectra with Uncertainty-Aware Machine Learning
• Fast and robust Bayesian Inference using Gaussian Processes with GPry
• Black-Box Optimization with Local Generative Surrogates
• Fisher Forecasts for Primordial non-Gaussianity from Persistent Homology
• Lightning-Fast Gravitational Wave Parameter Inference through Neural Amortization
• Primordial non-Gaussianity and non-Gaussian Covariance
• Bayesian model comparison for simulation-based inference
• Calibrating cosmological simulations with implicit likelihood inference using galaxy growth observables
• Parameter Estimation with Dense and Convolutional Neural Networks Applied to the FitzHugh-Nagumo ODE
• One never walks alone: the effect of the perturber population on subhalo measurements in strong gravitational lenses
• Differentiable Cosmological Simulation with Adjoint Method
• The emergence of scale-free fires in Australia
• Accurate X-ray Timing in the Presence of Systematic Biases With Simulation-Based Inference
• Some models are useful, but how do we know which ones? Towards a unified Bayesian model taxonomy
• Back to the Formula -- LHC Edition
• Strong-lensing source reconstruction with variationally optimised Gaussian processes
• Variational inference of fractional Brownian motion with linear computational complexity
• Multifidelity multilevel Monte Carlo to accelerate approximate Bayesian parameter inference for partially observed stochastic processes
• Uncovering dark matter density profiles in dwarf galaxies with graph neural networks
• Differentiable Matrix Elements with MadJax
• Approximate Bayesian Neural Doppler Imaging
• Bayesian evidence-driven likelihood selection for sky-averaged 21-cm signal extraction
• The Linear Template Fit
• Simple lessons from complex learning: what a neural network model learns about cosmic structure formation
• Fast and Flexible Analysis of Direct Dark Matter Search Data with Machine Learning
• Physics-Integrated Variational Autoencoders for Robust and Interpretable Generative Modeling
• Normalizing flows for random fields in cosmology
• Cosmology with cosmic web environments II. Redshift-space auto and cross power spectra
• Advancing Tools for Simulation-Based Inference
• Diagnostics for Conditional Density Models and Bayesian Inference Algorithms
• Dealing with Nuisance Parameters using Machine Learning in High Energy Physics: a Review
• Differentiable Likelihoods for Fast Inversion of 'Likelihood-Free' Dynamical Systems
• Neural Conditional Reweighting
• Particle Cloud Generation with Message Passing Generative Adversarial Networks
• Sequential Neural Posterior and Likelihood Approximation
• Simulation-based inference of single-molecule force spectroscopy
• Perturbation-theory informed integrators for cosmological simulations
• Data-Centric Engineering: integrating simulation, machine learning and statistics. Challenges and Opportunities
• Domain-informed neural networks for interaction localization within astroparticle experiments
• HNPE: Leveraging Global Parameters for Neural Posterior Estimation
• Confidence Sets and Hypothesis Testing in a Likelihood-Free Inference Setting
• Simulation Based Inference for Efficient Theory Space Sampling: an Application to Supersymmetric Explanations of the Anomalous Muon (g-2)
• Simulation-Based Inference with Approximately Correct Parameters via Maximum Entropy
• Approximate Bayesian Computation Applied to the Diffuse Gamma-ray Sky
• Real-time parameter inference in reduced-order flame models with heteroscedastic Bayesian neural network ensembles
• Gradient-based Bayesian Experimental Design for Implicit Models using Mutual Information Lower Bounds
• Analytic auto-differentiable $\Lambda$CDM cosmography
• Automating Inference of Binary Microlensing Events with Neural Density Estimation
• Amortized Bayesian Inference for Models of Cognition
• Teaching deep neural networks to localize single molecules for super-resolution microscopy
• Stratified sampling and bootstrapping for approximate Bayesian computation
• Discrepancy-based Inference for Intractable Generative Models using Quasi-Monte Carlo
• Approximate Bayesian inference from noisy likelihoods with Gaussian process emulated MCMC
• Approximate Bayesian Computation for an Explicit-Duration Hidden Markov Model of COVID-19 Hospital Trajectories
• Fitting summary statistics of neural data with a differentiable spiking network simulator
• Group equivariant neural posterior estimation
• A Note on Simulation-Based Inference by Matching Random Features
• Calibrating Over-Parametrized Simulation Models: A Framework via Eligibility Set
• Simulation-Based Inference for Global Health Decisions
• IV-Posterior: Inverse Value Estimation for Interpretable Policy Certificates
• ADAVI: Automatic Dual Amortized Variational Inference Applied To Pyramidal Bayesian Models
• A Geometric Perspective on Bayesian and Generalized Fiducial Inference
• MINIMALIST: Mutual INformatIon Maximization for Amortized Likelihood Inference from Sampled Trajectories
• Large Language Models -- the Future of Fundamental Physics?
• ABC-Di: Approximate Bayesian Computation for Discrete Data
• Deterministic particle flows for constraining stochastic nonlinear systems
• Optimal Design of Experiments for Simulation-Based Inference of Mechanistic Acyclic Biological Networks
• Population-based Optimization for Kinetic Parameter Identification in Glycolytic Pathway in Saccharomyces cerevisiae