Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows

2018/05/18 by George Papamakarios, Papamakarios, George, David C. Sterratt +3 · 63 citations
Computer Science · Mathematics · Physics and Astronomy · #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms #Markov Chains and Monte Carlo Methods #Model Reduction and Neural Networks #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1805.07226

Accepted for publication at AISTATS 2019

openalex publication_date 2018/05/18 · arxiv created 2019/01/21 · arxiv updated 2019/01/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior density. A sequential training procedure guides simulations and reduces simulation cost by orders of magnitude. We show that SNL is more robust, more accurate and requires less tuning than related neural-based methods, and we discuss diagnostics for assessing calibration, convergence and goodness-of-fit.

Cited by

• Inherited or produced? Inferring protein production kinetics when protein counts are shaped by a cell's division history
• Likelihood-Preserving Embeddings for Statistical Inference
• Minimum Distance Summaries for Robust Neural Posterior Estimation
• Global fits and the search for new physics: past, present and future
• Model-Free Assessment of Simulator Fidelity via Quantile Curves
• Assessment of Simulation-based Inference Methods for Stochastic Compartmental Models
• Simulation and inference methods for non-Markovian stochastic biochemical reaction networks
• MIST: Mutual Information Estimation Via Supervised Training
• Fast and Robust Simulation-Based Inference With Optimization Monte Carlo
• On the transmission of texts: Written cultures as complex systems
• A Comparison of Kernels for ABC-SMC
• Dark Energy Survey Year 3 results: Simulation-based wCDM inference from weak lensing and galaxy clustering maps with deep learning: Analysis design
• ASBI: Leveraging Informative Real-World Data for Active Black-Box Simulator Tuning
• SBi3PCF: Simulation-based inference with the integrated 3PCF
• PriorGuide: Test-Time Prior Adaptation for Simulation-Based Inference
• Simulation-based inference at the theoretical limit for fast, robust microstructural MRI with minimal diffusion data
• Simulation-based Inference of Massive Black Hole Binaries using Sequential Neural Likelihood
• CP4SBI: Local Conformal Calibration of Credible Sets in Simulation-Based Inference
• Misspecification-robust amortised simulation-based inference using variational methods
• Cosmo-Learn: code for learning cosmology using different methods and mock data
• Mining gold from implicit models to improve likelihood-free inference
• Detecting conflicting summary statistics in likelihood-free inference
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• Arbitrary Marginal Neural Ratio Estimation for Simulation-based Inference
• Cosmic string gravitational wave backgrounds at LISA: I. Signal survey, template reconstruction, and model comparison
• Quantitative and Predictive Folding Models from Limited Single-Molecule Data Using Simulation-Based Inference
• Hamiltonian parameter inference from resonant inelastic x-ray scattering with active learning
• Simulation-based inference for Precision Neutrino Physics through Neural Monte Carlo tuning
• How to embed any likelihood into SBI: Application to Planck + Stage IV galaxy surveys and Dynamical Dark Energy
• Flows for simultaneous manifold learning and density estimation
• Graphical Normalizing Flows
• Recent Advances in Simulation-based Inference for Gravitational Wave Data Analysis
• Estimating Marginal Likelihoods in Likelihood-Free Inference via Neural Density Estimation
• ABC-Di: Approximate Bayesian Computation for Discrete Data
• Simulation-Efficient Cosmological Inference with Multi-Fidelity SBI
• Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference
• Simulation-based population inference of LISA's Galactic binaries: Bypassing the global fit
• Inverse Gaussian Process regression for likelihood-free inference
• Direct Fisher Score Estimation for Likelihood Maximization
• Multilevel neural simulation-based inference
• Variational Transdimensional Inference
• On the Need to Align Intent and Implementation in Uncertainty Quantification for Machine Learning
• Flexible Selective Inference with Flow-based Transport Maps
• A thermodynamic approach to Approximate Bayesian Computation with multiple summary statistics
• FNOPE: Simulation-based inference on function spaces with Fourier Neural Operators
• Causal Posterior Estimation
• Probabilistic Software Modeling: A Data-driven Paradigm for Software Analysis
• Simulation-based Bayesian inference for multi-fingered robotic grasping
• Black-Box Inference for Non-Linear Latent Force Models
• ConDiSim: Conditional Diffusion Models for Simulation Based Inference
• Uncertainty-Aware Surrogate-based Amortized Bayesian Inference for Computationally Expensive Models
• Robust Simulation Based Inference Through Robust Optimal Transport
• BayesFlow 2: Multi-Backend Amortized Bayesian Inference in Python
• Amortized mean-shift interacting particles
• Neural Density Estimation and Likelihood-free Inference
• A multifidelity approximate Bayesian computation with pre-filtering
• Neural Posterior Estimation for Stochastic Epidemic Models Using Final Outcome Data
• Truncated Neural Likelihood Estimation for Simulation-Based Inference in State-Space Models
• Real-time Approximate Bayesian Computation for Scene Understanding
• Effortless, Simulation-Efficient Bayesian Inference using Tabular Foundation Models
• A guide to constraining effective field theories with machine learning
• Reducing Calls to the Simulator in Simulation Based Inference (SBI)
• Simulation-based inference for stochastic nonlinear mixed-effects models with applications in systems biology

Related

• Sequential Neural Methods for Likelihood-free Inference
• Sequential Neural Posterior and Likelihood Approximation
• An efficient likelihood-free Bayesian inference method based on sequential neural posterior estimation
• Misspecification-robust Sequential Neural Likelihood for Simulation-based Inference
• Sequential Neural Score Estimation: Likelihood-Free Inference with Conditional Score Based Diffusion Models