Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks

Autor:	Yujun Teng, Paris Perdikaris, Sifan Wang
Rok vydání:	2021
Předmět:	Computational Mathematics Conservation law Theoretical computer science Artificial neural network Differential equation business.industry Applied Mathematics Deep learning Domain knowledge Artificial intelligence Balanced flow business Mathematics
Zdroj:	SIAM Journal on Scientific Computing. 43:A3055-A3081
ISSN:	1095-7197 1064-8275
DOI:	10.1137/20m1318043
Popis:	The widespread use of neural networks across different scientific domains often involves constraining them to satisfy certain symmetries, conservation laws, or other domain knowledge. Such constrai...
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::1026f9a74fe5ed0c6e3c170da0a0b52e https://doi.org/10.1137/20m1318043 Zobrazit plný text záznamu