Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ye, Zhanhong, Huang, Xiang, Chen, Leheng, Liu, Zining, Wu, Bingyang, Liu, Hongsheng, Wang, Zidong, Dong, Bin
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2407.06664
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912204821889024
author Ye, Zhanhong
Huang, Xiang
Chen, Leheng
Liu, Zining
Wu, Bingyang
Liu, Hongsheng
Wang, Zidong
Dong, Bin
author_facet Ye, Zhanhong
Huang, Xiang
Chen, Leheng
Liu, Zining
Wu, Bingyang
Liu, Hongsheng
Wang, Zidong
Dong, Bin
contents This paper introduces PDEformer-1, a versatile neural solver capable of simultaneously addressing various partial differential equations (PDEs). With the PDE represented as a computational graph, we facilitate the seamless integration of symbolic and numeric information inherent in a PDE. A graph Transformer and an implicit neural representation (INR) are employed subsequently to generate mesh-free predicted solutions. We generated a dataset with up to three million samples involving diverse one-dimensional PDEs to pretrain our model. Compared with baseline models trained specifically on benchmark datasets, our pretrained model achieves comparable accuracy via zero-shot inference, and the advantage expands after finetuning. For PDEs new or unseen in the pretraining stage, our model can adapt quickly by finetuning on a relatively small set of examples from the target equation. Additionally, PDEformer-1 demonstrates promising results in the inverse problem of PDE scalar coefficient recovery and coefficient field recovery.
format Preprint
id arxiv_https___arxiv_org_abs_2407_06664
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle PDEformer-1: A Foundation Model for One-Dimensional Partial Differential Equations
Ye, Zhanhong
Huang, Xiang
Chen, Leheng
Liu, Zining
Wu, Bingyang
Liu, Hongsheng
Wang, Zidong
Dong, Bin
Numerical Analysis
This paper introduces PDEformer-1, a versatile neural solver capable of simultaneously addressing various partial differential equations (PDEs). With the PDE represented as a computational graph, we facilitate the seamless integration of symbolic and numeric information inherent in a PDE. A graph Transformer and an implicit neural representation (INR) are employed subsequently to generate mesh-free predicted solutions. We generated a dataset with up to three million samples involving diverse one-dimensional PDEs to pretrain our model. Compared with baseline models trained specifically on benchmark datasets, our pretrained model achieves comparable accuracy via zero-shot inference, and the advantage expands after finetuning. For PDEs new or unseen in the pretraining stage, our model can adapt quickly by finetuning on a relatively small set of examples from the target equation. Additionally, PDEformer-1 demonstrates promising results in the inverse problem of PDE scalar coefficient recovery and coefficient field recovery.
title PDEformer-1: A Foundation Model for One-Dimensional Partial Differential Equations
topic Numerical Analysis
url https://arxiv.org/abs/2407.06664