OmniArch: Building Foundation Model For Scientific Computing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Tianyu, Zhou, Haoyi, Li, Ying, Wang, Hao, Gao, Chonghan, Shi, Rongye, Zhang, Shanghang, Li, Jianxin
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909626278084608
author Chen, Tianyu
Zhou, Haoyi
Li, Ying
Wang, Hao
Gao, Chonghan
Shi, Rongye
Zhang, Shanghang
Li, Jianxin
author_facet Chen, Tianyu
Zhou, Haoyi
Li, Ying
Wang, Hao
Gao, Chonghan
Shi, Rongye
Zhang, Shanghang
Li, Jianxin
contents Foundation models have revolutionized language modeling, while whether this success is replicated in scientific computing remains unexplored. We present OmniArch, the first prototype aiming at solving multi-scale and multi-physics scientific computing problems with physical alignment. We addressed all three challenges with one unified architecture. Its pre-training stage contains a Fourier Encoder-decoder fading out the disharmony across separated dimensions and a Transformer backbone integrating quantities through temporal dynamics, and the novel PDE-Aligner performs physics-informed fine-tuning under flexible conditions. As far as we know, we first conduct 1D-2D-3D united pre-training on the PDEBench, and it sets not only new performance benchmarks for 1D, 2D, and 3D PDEs but also demonstrates exceptional adaptability to new physics via in-context and zero-shot learning approaches, which supports realistic engineering applications and foresight physics discovery.
format Preprint
id arxiv_https___arxiv_org_abs_2402_16014
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle OmniArch: Building Foundation Model For Scientific Computing
Chen, Tianyu
Zhou, Haoyi
Li, Ying
Wang, Hao
Gao, Chonghan
Shi, Rongye
Zhang, Shanghang
Li, Jianxin
Machine Learning
Artificial Intelligence
Foundation models have revolutionized language modeling, while whether this success is replicated in scientific computing remains unexplored. We present OmniArch, the first prototype aiming at solving multi-scale and multi-physics scientific computing problems with physical alignment. We addressed all three challenges with one unified architecture. Its pre-training stage contains a Fourier Encoder-decoder fading out the disharmony across separated dimensions and a Transformer backbone integrating quantities through temporal dynamics, and the novel PDE-Aligner performs physics-informed fine-tuning under flexible conditions. As far as we know, we first conduct 1D-2D-3D united pre-training on the PDEBench, and it sets not only new performance benchmarks for 1D, 2D, and 3D PDEs but also demonstrates exceptional adaptability to new physics via in-context and zero-shot learning approaches, which supports realistic engineering applications and foresight physics discovery.
title OmniArch: Building Foundation Model For Scientific Computing
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2402.16014