Neural Conditional Simulation for Complex Spatial Processes

2025-08-27

AI Summary

This paper addresses the challenge of conditional simulation for complex spatial processes at unobserved locations—particularly when conventional methods (e.g., MCMC) are computationally prohibitive or intractable to sample from. We propose Neural Conditional Simulation (NCS), a novel framework built upon score-based diffusion models with spatial masking—the first application of neural diffusion models to spatial conditional simulation. Crucially, NCS is trained solely on unconditional samples and achieves zero-shot generalization to arbitrary observation configurations and model parameters without fine-tuning or retraining. Evaluated on Gaussian processes and the Brown–Resnick max-stable process, NCS faithfully reproduces the true predictive distribution, preserving statistical consistency. Moreover, it accelerates simulation by over an order of magnitude relative to MCMC while achieving higher accuracy. These advances significantly enhance both the efficiency and scalability of spatial prediction and uncertainty quantification.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Sampling/Simulation-based Search
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Web Mining and Content Analysis: Web data generation and simulation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

A key objective in spatial statistics is to simulate from the distribution of a spatial process at a selection of unobserved locations conditional on observations (i.e., a predictive distribution) to enable spatial prediction and uncertainty quantification. However, exact conditional simulation from this predictive distribution is intractable or inefficient for many spatial process models. In this paper, we propose neural conditional simulation (NCS), a general method for spatial conditional simulation that is based on neural diffusion models. Specifically, using spatial masks, we implement a conditional score-based diffusion model that evolves Gaussian noise into samples from a predictive distribution when given a partially observed spatial field and spatial process parameters as inputs. The diffusion model relies on a neural network that only requires unconditional samples from the spatial process for training. Once trained, the diffusion model is amortized with respect to the observations in the partially observed field, the number and locations of those observations, and the spatial process parameters, and can therefore be used to conditionally simulate from a broad class of predictive distributions without retraining the neural network. We assess the NCS-generated simulations against simulations from the true conditional distribution of a Gaussian process model, and against Markov chain Monte Carlo (MCMC) simulations from a Brown--Resnick process model for spatial extremes. In the latter case, we show that it is more efficient and accurate to conditionally simulate using NCS than classical MCMC techniques implemented in standard software. We conclude that NCS enables efficient and accurate conditional simulation from spatial predictive distributions that are challenging to sample from using traditional methods.

Problem

• Simulating conditional spatial processes for prediction and uncertainty quantification
• Overcoming intractability of exact conditional simulation in spatial models
• Providing efficient and accurate sampling from complex predictive distributions

Innovation

• Neural diffusion models for spatial conditional simulation
• Amortized neural network trained with unconditional samples
• Efficient conditional simulation without retraining neural network