Accelerating Training of Physics Informed Neural Network for 1D PDEs with Hierarchical Matrices.
Mateusz DobijaAnna PaszynskaCarlos UriarteMaciej PaszynskiPublished in: ICCS (3) (2024)
Keyphrases
- neural network
- training algorithm
- training process
- multi layer perceptron
- feed forward neural networks
- feedforward neural networks
- neural network training
- training patterns
- back propagation
- computer science
- backpropagation algorithm
- partial differential equations
- artificial neural networks
- level set
- adaptive learning rate
- neural network model
- train a neural network
- fault diagnosis
- training examples
- online learning
- fuzzy logic
- training set
- pattern recognition
- image processing
- artificial intelligence
- training phase
- neural network is trained
- error back propagation
- multilayer neural network
- positive definite
- covariance matrices
- feed forward
- hidden layer
- learning rate
- multi layer
- neural nets
- multilayer perceptron
- singular value decomposition
- image denoising
- hierarchical structure
- self organizing maps
- test set
- multiscale
- genetic algorithm