A failure-informed multi-stage training algorithm for three-component nonlinear Schrödinger equation.
Yawen WuLiming LingYubin HuangPublished in: Comput. Math. Appl. (2024)
Keyphrases
- multistage
- training algorithm
- wave equation
- hamilton jacobi
- finite difference
- back propagation
- heat equation
- neural network
- neural network training
- dynamic programming
- single stage
- lot sizing
- support vector machine
- optimal policy
- partial differential equations
- learning rate
- eikonal equation
- medial axis
- learning algorithm
- rbf neural network
- numerical solution
- training process
- diffusion process
- hidden layer
- scale spaces
- rbf network
- machine learning
- training data