FNOPE: Simulation-based inference on function spaces with Fourier Neural Operators

2025/05/28 by Guy Moss, Leah Sophie Muhle, Moss, Guy +7
Computer Science · Physics and Astronomy · #Gaussian Processes and Bayesian Inference #Generative Adversarial Networks and Image Synthesis #Model Reduction and Neural Networks

paper · pdf · doi:10.48550/arxiv.2505.22573

Abstract

Simulation-based inference (SBI) is an established approach for performing Bayesian inference on scientific simulators. SBI so far works best on low-dimensional parametric models. However, it is difficult to infer function-valued parameters, which frequently occur in disciplines that model spatiotemporal processes such as the climate and earth sciences. Here, we introduce an approach for efficient posterior estimation, using a Fourier Neural Operator (FNO) architecture with a flow matching objective. We show that our approach, FNOPE, can perform inference of function-valued parameters at a fraction of the simulation budget of state of the art methods. In addition, FNOPE supports posterior evaluation at arbitrary discretizations of the domain, as well as simultaneous estimation of vector-valued parameters. We demonstrate the effectiveness of our approach on several benchmark tasks and a challenging spatial inference task from glaciology. FNOPE extends the applicability of SBI methods to new scientific domains by enabling the inference of function-valued parameters.

Citations

• Sensitivity-Constrained Fourier Neural Operators for Forward and Inverse Problems in Parametric Differential Equations
• sbi reloaded: a toolkit for simulation-based inference workflows
• Amortized Probabilistic Conditioning for Optimization, Simulation and Inference
• Latent Neural Operator for Solving Forward and Inverse PDE Problems
• Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priors
• FUSE: Fast Unified Simulation and Estimation for PDEs
• All-in-one simulation-based inference
• Diffusion posterior sampling for simulation-based inference in tall data settings
• Invertible Fourier Neural Operators for Tackling Both Forward and Inverse Problems
• Transolver: A Fast Transformer Solver for PDEs on General Geometries
• Simulation-based inference of surface accumulation and basal melt rates of an Antarctic ice shelf from isochronal layers
• Geometry-Informed Neural Operator for Large-Scale 3D PDEs
• Idealized Baroclinic Wave Test Results
• Beyond Regular Grids: Fourier-Based Neural Operators on Arbitrary Domains
• Flow Matching for Scalable Simulation-Based Inference
• GNOT: A General Neural Operator Transformer for Operator Learning
• Variational Autoencoding Neural Operators
• Score-based Diffusion Models in Function Space
• Improving and generalizing flow-based generative models with minibatch optimal transport
• Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model
• Modeling Temporal Data as Continuous Functions with Stochastic Process Diffusion
• Flow Matching for Generative Modeling
• Diffusion Posterior Sampling for General Noisy Inverse Problems
• Compositional Score Modeling for Simulation-based Inference
• Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow
• Generative Adversarial Neural Operators
• GATSBI: Generative Adversarial Training for Simulation-Based Inference
• Denoising Diffusion Restoration Models
• Physics-Informed Neural Operator for Learning Partial Differential Equations
• A Probabilistic State Space Model for Joint Inference from Differential Equations and Data
• Benchmarking Simulation-Based Inference
• Score-Based Generative Modeling through Stochastic Differential Equations
• Fourier Neural Operator for Parametric Partial Differential Equations
• BayesFlow: Learning complex stochastic models with invertible neural networks
• On Contrastive Learning for Likelihood-free Inference
• Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators
• Neural Spline Flows
• Automatic Posterior Transformation for Likelihood-Free Inference
• Forward Amortized Inference for Likelihood-Free Variational Marginalization
• Sequential Neural Likelihood: Fast Likelihood-free Inference with Autoregressive Flows
• Validating Bayesian Inference Algorithms with Simulation-Based Calibration
• Flexible statistical inference for mechanistic models of neural dynamics
• Fast ε-free Inference of Simulation Models with Bayesian Conditional Density Estimation
• Geostatistical inference of hydraulic conductivity and dispersivities from hydraulic heads and tracer data

Related

• Robust Ocean Subgrid-Scale Parameterizations Using Fourier Neural Operators
• From Fourier to Neural ODEs: Flow Matching for Modeling Complex Systems
• Flow Matching for Scalable Simulation-Based Inference
• Fourier Neural Operators for Arbitrary Resolution Climate Data Downscaling
• Idealized Baroclinic Wave Test Results