Share your thoughts, 1 month free Claude Pro on usSee more
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
PDE solvingPoisson
Time (s)802
55
Continuum Field Reconstruction (Rollout)2D Navier-Stokes nu=1e-3
MSE1.339
54
Forward PDE solvingHelmholtz
Relative Error0.088
26
Solving PDEBurgers
Relative Error8.61
24
Inverse ProblemPoisson
PDE Residual178.5
21
Continuum Field Reconstruction (Rollout)2D Navier-Stokes nu=1e-5
MSE1.302
18
Continuum Field ReconstructionShallow-Water (In-t)
MSE2.734
18
Continuum Field ReconstructionShallow-Water (Out-t)
MSE3.617
18
Continuum Field ReconstructionShallow-Water Avg
MSE3.175
18
Continuum Field ReconstructionNSν1e-5 Avg
MSE1.094
18
Showing 10 of 139 rows
...

Other info

Code

Follow for update