Fast $\epsilon$-free Inference of Simulation Models with Bayesian Conditional Density Estimation

Abstract

Many statistical models can be simulated forwards but have intractable likelihoods. Approximate Bayesian Computation (ABC) methods are used to infer properties of these models from data. Traditionally these methods approximate the posterior over parameters by conditioning on data being inside an $\epsilon$-ball around the observed data, which is only correct in the limit $\epsilon \to 0$. Monte Carlo methods can then draw samples from the approximate posterior to approximate predictions or error bars on parameters. These algorithms critically slow down as $\epsilon \to 0$, and in practice draw samples from a broader distribution than the posterior. We propose a new approach to likelihood-free inference based on Bayesian conditional density estimation. Preliminary inferences based on limited simulation data are used to guide later simulations. In some cases, learning an accurate parametric representation of the entire true posterior distribution requires fewer model simulations than Monte Carlo ABC methods need to produce a single sample from an approximate posterior.

Appeared at NIPS 2016. Fixed typo in Eq (37)

Cited by in corpus (58)

• The frontier of simulation-based inference
• Masked Autoregressive Flow for Density Estimation
• Mining gold from implicit models to improve likelihood-free inference
• Hierarchical Implicit Models and Likelihood-Free Variational Inference
• Bayesian optimisation for likelihood-free cosmological inference
• Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows
• Fast and Credible Likelihood-Free Cosmology with Truncated Marginal Neural Ratio Estimation
• Conditional Density Estimation Tools in Python and R with Applications to Photometric Redshifts and Likelihood-Free Cosmological Inference
• A neural simulation-based inference approach for characterizing the Galactic Center $\gamma$-ray excess
• A Likelihood-Free Inference Framework for Population Genetic Data using Exchangeable Neural Networks
• Conditional Density Estimation with Bayesian Normalising Flows
• Efficient acquisition rules for model-based approximate Bayesian computation
• ABC-CDE: Towards Approximate Bayesian Computation with Complex High-Dimensional Data and Limited Simulations
• The sum of the masses of the Milky Way and M31: a likelihood-free inference approach
• Simulation-based inference of dynamical galaxy cluster masses with 3D convolutional neural networks
• Variational Bayesian Monte Carlo with Noisy Likelihoods
• Likelihood-free MCMC with Amortized Approximate Ratio Estimators
• Neural parameters estimation for brain tumor growth modeling
• Applications of machine learning in gravitational wave research with current interferometric detectors
• Effective LHC measurements with matrix elements and machine learning
• Neural Approximate Sufficient Statistics for Implicit Models
• Benchmarking Simulation-Based Inference
• One never walks alone: the effect of the perturber population on subhalo measurements in strong gravitational lenses
• Likelihood-free inference with an improved cross-entropy estimator
• Accurate X-ray Timing in the Presence of Systematic Biases With Simulation-Based Inference
• Cosmological Parameter Estimation and Inference using Deep Summaries
• Fast likelihood-free inference in the LSS Stage IV era
• Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference
• Variational Implicit Processes
• Batch simulations and uncertainty quantification in Gaussian process surrogate approximate Bayesian computation
• Neural Density Estimation and Likelihood-free Inference
• Trial by FIRE: Probing the dark matter density profile of dwarf galaxies with GraphNPE
• Sequential Neural Posterior and Likelihood Approximation
• Sequential Neural Methods for Likelihood-free Inference
• HNPE: Leveraging Global Parameters for Neural Posterior Estimation
• Approximate Bayesian Computation via Classification
• Likelihood-free stochastic approximation EM for inference in complex models
• Gaussian process modeling in approximate Bayesian computation to estimate horizontal gene transfer in bacteria
• Stratified sampling and bootstrapping for approximate Bayesian computation
• Robust Optimisation Monte Carlo
• Estimating Nonlinear Dynamics with the ConvNet Smoother
• Distortion estimates for approximate Bayesian inference
• Simulation-based Bayesian inference for multi-fingered robotic grasping
• Cosmic Strings-induced CMB anisotropies in light of Weighted Morphology
• Fitting summary statistics of neural data with a differentiable spiking network simulator
• Group equivariant neural posterior estimation
• A review of Approximate Bayesian Computation methods via density estimation: inference for simulator-models
• Semi-Implicit Neural Solver for Time-dependent Partial Differential Equations
• Variational Inference over Non-differentiable Cardiac Simulators using Bayesian Optimization
• Easy High-Dimensional Likelihood-Free Inference
• Bayesian Deconditional Kernel Mean Embeddings
• MINIMALIST: Mutual INformatIon Maximization for Amortized Likelihood Inference from Sampled Trajectories
• STReSSD: Sim-To-Real from Sound for Stochastic Dynamics
• Real-time Approximate Bayesian Computation for Scene Understanding
• Detecting conflicting summary statistics in likelihood-free inference
• Optimal Design of Experiments for Simulation-Based Inference of Mechanistic Acyclic Biological Networks
• Inverse Gaussian Process regression for likelihood-free inference
• IV-Posterior: Inverse Value Estimation for Interpretable Policy Certificates