Computational Science
物理信息Kolmogorov–Arnold网络
Physics-informed Kolmogorov–Arnold NetworksA deep learning framework that integrates physical laws and governing equations into the training process of Kolmogorov–Arnold Networks (KANs) to improve prediction accuracy and generalizability. / 一种深度学习框架，将物理定律和控制方程集成到Kolmogorov–Arnold网络（KAN）的训练过程中，以提高预测精度和泛化能力。
- Knowledge type
- 工艺技术
- First proposed or discovered
- Proposer
- Turkey Farm member
- Creator
- 农场主
- Last updated
- 2026-10-31 16:06:38
Sub-knowledge points
Kolmogorov–Arnold representation theoremB-spline activation functionsPhysics-informed neural networks (PINNs) methodology
Higher-order knowledge points
Main parameters and performance indicators
Physics-informed loss function: Integrates governing equations into the optimization objectiveKolmogorov–Arnold Network architecture: Multi-layer perceptron with B-spline activation functions
Open problems
Optimizing the balance between data-driven and physics-driven termsScalability to high-dimensional systems
Related literature
ranked by community starsFatigue crack growth remaining life prediction via physics-informed Kolmogorov–Arnold NetworksOpen source
Related products and links
No item recorded yet
Creation and modification trail
1 recordsRecord title: Physics-informed Kolmogorov–Arnold Networks.
Node update feedback
review required