DeepONet as a multi-operator extrapolation model: distributed pretraining with physics-informed fine-tuning

DOI10.1016/J.JCP.2025.114537MaRDI QIDQ6853389zbMATH OpenFDO

Authors Zecheng Zhang, Christian Moya, Lu Lu, Guang Lin, Hayden Schaeffer

Publication date 25 February 2026

Published in Journal of Computational Physics (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1016/J.JCP.2025.114537

zbMATH Keywords

fine-tuningpretrainingphysics-informed learningdistributed trainingdeep operator learninglow-rank adaptation (lora)

Mathematics Subject Classification ID

Artificial intelligence (68Txx) Numerical methods for partial differential equations, boundary value problems (65Nxx) Numerical methods for partial differential equations, initial value and time-dependent initial-boundary value problems (65Mxx)

Cites work

• A comprehensive and fair comparison of two neural operators (with practical extensions) based on FAIR data
• A physics-informed variational DeepONet for predicting crack path in quasi-brittle materials
• B-DeepONet: an enhanced Bayesian deeponet for solving noisy parametric PDEs using accelerated replica exchange SGLD
• Bayesian deep operator learning for homogenized to fine-scale maps for multiscale PDE
• Convergence rate of DeepONets for learning operators arising from advection-diffusion equations
• D2NO: efficient handling of heterogeneous input function spaces with distributed deep neural operators
• DeepM\&Mnet for hypersonics: predicting the coupled flow and finite-rate chemistry behind a normal shock using neural-network approximation of operators
• DeepM\&Mnet: inferring the electroconvection multiphysics fields based on operator approximation by neural networks
• DeepXDE: a deep learning library for solving differential equations
• Fourier-DeepONet: Fourier-enhanced deep operator networks for full waveform inversion with improved accuracy, generalizability, and robustness
• Learning partial differential equations via data discovery and sparse optimization
• MIONet: Learning Multiple-Input Operators via Tensor Product
• MODNO: multi-operator learning with distributed neural operators
• NH-PINN: neural homogenization-based physics-informed neural network for multiscale problems
• Neural operator prediction of linear instability waves in high-speed boundary layers
• Reliable extrapolation of deep neural operators informed by physics or sparse observations

This page was built for publication: DeepONet as a multi-operator extrapolation model: distributed pretraining with physics-informed fine-tuning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6853389)