Who breaks early, looses: goal oriented training of deep neural networks based on port Hamiltonian dynamics.
Julian BurghoffMarc Heinrich MonellsHanno GottschalkPublished in: CoRR (2023)
Keyphrases
- goal oriented
- neural network
- recurrent networks
- training process
- training algorithm
- back propagation
- multi layer perceptron
- backpropagation algorithm
- feed forward neural networks
- pattern recognition
- requirements analysis
- requirements engineering
- neural network structure
- process oriented
- feedforward neural networks
- recurrent neural networks
- fuzzy logic
- deep architectures
- error back propagation
- neural network model
- training examples
- training samples
- multi layer
- dynamical systems
- fault diagnosis
- test set
- data marts
- case study
- learning algorithm
- dynamic model
- radial basis function
- business processes
- artificial neural networks
- training set
- information systems