Abstract:The accessibility of spatially distributed data, enabled by affordable sensors, field, and numerical experiments, has facilitated the development of data-driven solutions for scientific problems, including climate change, weather prediction, and urban planning. Neural Partial Differential Equations (Neural PDEs), which combine deep learning (DL) techniques with domain expertise (e.g., governing equations) for parameterization, have proven to be effective in capturing valuable correlations within spatiotemporal datasets. However, sparse and noisy measurements coupled with modeling approximation introduce aleatoric and epistemic uncertainties. Therefore, quantifying uncertainties propagated from model inputs to outputs remains a challenge and an essential goal for establishing the trustworthiness of Neural PDEs. This work evaluates various Uncertainty Quantification (UQ) approaches for both Forward and Inverse Problems in scientific applications. Specifically, we investigate the effectiveness of Bayesian methods, such as Hamiltonian Monte Carlo (HMC) and Monte-Carlo Dropout (MCD), and a more conventional approach, Deep Ensembles (DE). To illustrate their performance, we take two canonical PDEs: Burger's equation and the Navier-Stokes equation. Our results indicate that Neural PDEs can effectively reconstruct flow systems and predict the associated unknown parameters. However, it is noteworthy that the results derived from Bayesian methods, based on our observations, tend to display a higher degree of certainty in their predictions as compared to those obtained using the DE. This elevated certainty in predictions suggests that Bayesian techniques might underestimate the true underlying uncertainty, thereby appearing more confident in their predictions than the DE approach.

Solution method and parameter estimation of uncertain partial differential equation with application to China's population

Residual analysis and parameter estimation of uncertain differential equations

Parameter estimation in uncertain differential equations with exponential solutions

Moment estimation for parameters in high-order uncertain differential equations

Parameter estimation for uncertain fractional differential equations

Uncertain partial differential equation with application to heat conduction

On weighted threshold moment estimation of uncertain differential equations with applications in interbank rates analysis

Least squares parameter estimation for uncertain fractional differential equations and application to stock model

Initial Value Estimation of Uncertain Differential Equations and Zero-Day of COVID-19 Spread in China.

Estimating Unknown Time-Varying Parameters in Uncertain Differential Equation

The inverse uncertainty distribution of the solutions to a class of higher-order uncertain differential equations

Estimating Unknown Parameters in Uncertain Differential Equation by Maximum Likelihood Estimation

Numerical solution and parameter estimation for uncertain SIR model with application to COVID-19

High-order Uncertain Differential Equation and Its Application to Nuclear Reactors

Parameter Estimation of Uncertain Differential Equations Driven by Threshold Ornstein–Uhlenbeck Process with Application to U.S. Treasury Rate Analysis

Bayesian Numerical Methods for Nonlinear Partial Differential Equations

Parameter Estimation for Multivariate Nonlinear Stochastic Differential Equation Models: A Comparison Study

DEEP LEARNING OF PARAMETERIZED EQUATIONS WITH APPLICATIONS TO UNCERTAINTY QUANTIFICATION

Evaluating Uncertainty Quantification approaches for Neural PDEs in scientific applications

A Low-Rank Solver for Parameter Estimation and Uncertainty Quantification in Time-Dependent Systems of Partial Differential Equations

Uncertainty Quantification for Forward and Inverse Problems of PDEs via Latent Global Evolution