Inverse Dirichlet weighting enables reliable training of physics informed neural networks.
Suryanarayana MadduDominik SturmChristian L. MüllerIvo F. SbalzariniPublished in: Mach. Learn. Sci. Technol. (2022)
Keyphrases
- neural network
- training process
- feedforward neural networks
- training algorithm
- artificial neural networks
- feed forward neural networks
- pattern recognition
- multi layer perceptron
- computer science
- radial basis function network
- neural network training
- back propagation
- training patterns
- cost effective
- backpropagation algorithm
- boundary conditions
- feed forward
- neural network model
- error back propagation
- artificial intelligence
- training set
- fuzzy logic
- genetic algorithm
- self organizing maps
- neural network structure
- support vector
- supervised learning
- training samples
- multi layer
- multilayer perceptron
- test set
- tf idf
- hidden markov models
- active learning
- training examples
- image segmentation