Conformalized Physics-Informed Neural Networks

Lena Podina, Mahdi Torabi Rad, Mohammad Kohandel
5/13/2024
cs.LG

No official code found. Best implementation: jayroxis/PINNs (722 stars, last commit 2024-05-20).

Abstract

Physics-informed neural networks (PINNs) are an influential method of solving differential equations and estimating their parameters given data. However, since they make use of neural networks, they provide only a point estimate of differential equation parameters, as well as the solution at any given point, without any measure of uncertainty. Ensemble and Bayesian methods have been previously applied to quantify the uncertainty of PINNs, but these methods may require making strong assumptions on the data-generating process, and can be computationally expensive. Here, we introduce Conformalized PINNs (C-PINNs) that, without making any additional assumptions, utilize the framework of conformal prediction to quantify the uncertainty of PINNs by providing intervals that have finite-sample, distribution-free statistical validity.

View on arXivView PDF

Code Implementations(1)

jayroxis/PINNsCommunity100%

PyTorch Implementation of Physics-informed Neural Networks

722182Python, Jupyter NotebookJan 30, 20202 years ago
pinnspytorch-implementation

Similar Papers

On the Generalization of PINNs outside the training domain and the Hyperparameters influencing it

85% match2023

Improved Uncertainty Quantification in Physics-Informed Neural Networks Using Error Bounds and Solution Bundles

85% match1 repo32025

Physics-Informed Neural Networks and Extensions

85% match2024

Uncertainty Quantification for Physics-Informed Neural Networks with Extended Fiducial Inference

84% match2025

Neural Measures for learning distributions of Random PDEs

84% match2025

THINNs: Thermodynamically Informed Neural Networks

84% match2025

Cite this paper

@article{podina2024conformalized,
  title  = {Conformalized Physics-Informed Neural Networks},
  author = {Lena Podina and Mahdi Torabi Rad and Mohammad Kohandel},
  year   = {2024},
  eprint = {2405.08111},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2405.08111v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2405.08111.svg)](https://codeofpaper.com/paper/2405.08111)

Discussion