A conservative hybrid physics-informed neural network method for Maxwell-Ampère-Nernst-Planck equations.
Cheng ChangZhouping XinTieyong ZengPublished in: CoRR (2023)
Keyphrases
- neural network
- pairwise
- preprocessing
- synthetic data
- cost function
- similarity measure
- artificial neural networks
- network model
- neural nets
- bp neural network
- high accuracy
- high precision
- auto associative
- training process
- detection method
- model selection
- edge detection
- computational cost
- significant improvement
- prior knowledge
- objective function
- decision trees