Chaos (Woodbury, N.Y.)
Multiprecision computing for multistage fractional physics-informed neural networks.
Na Xue, Minghua Chen
Published: 202510.1063/5.0287809
Abstract
Fractional physics-informed neural networks (fPINNs) have been successfully introduced in Pang et al. [SIAM J. Sci. Comput. 41, A2603-A2626 (2019)], which observe relative errors of 10-3∼10-4 for the subdiffusion equations. However, their high-precis…
Preview only. Read the full abstract at the source