Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

DiffusionPDE: Generative PDE-Solving Under Partial Observation

About

We introduce a general framework for solving partial differential equations (PDEs) using generative diffusion models. In particular, we focus on the scenarios where we do not have the full knowledge of the scene necessary to apply classical solvers. Most existing forward or inverse PDE approaches perform poorly when the observations on the data or the underlying coefficients are incomplete, which is a common assumption for real-world measurements. In this work, we propose DiffusionPDE that can simultaneously fill in the missing information and solve a PDE by modeling the joint distribution of the solution and coefficient spaces. We show that the learned generative priors lead to a versatile framework for accurately solving a wide range of PDEs under partial observation, significantly outperforming the state-of-the-art methods for both forward and inverse directions.

Jiahe Huang, Guandao Yang, Zichen Wang, Jeong Joon Park• 2024

Related benchmarks

TaskDatasetResultRank
Inverse ProblemPoisson
PDE Residual178.5
21
Air pollution forecastingChangshu Mobile
MAE8.77
17
Air pollution forecastingChangshu National
MAE8.81
17
Air pollution forecastingNanjing National
MAE7.78
17
Air pollution forecastingNanjing Mobile
MAE10.2
17
PM2.5 forecastingChangshu Mobile (test)
Training Time (s)124.5
15
PM2.5 forecastingNanjing Mobile
Training Time (s)155
15
PM2.5 forecastingChangshu (National) (test)
Training Time (s)146.3
15
PM2.5 forecastingNanjing (National) (test)
Training Time (s)151.6
15
Pollution Alert PredictionChangshu National
Recall0.5
15
Showing 10 of 108 rows
...

Other info

Code

Follow for update