Layer-Wise Interpretation of Deep Neural Networks Using Identity Initialization.
Shohei KubotaHideaki HayashiTomohiro HayaseSeiichi UchidaPublished in: CoRR (2021)
Keyphrases
- neural network
- multi layer
- multiple layers
- single layer
- pattern recognition
- artificial neural networks
- neural nets
- feed forward
- back propagation
- feed forward neural networks
- pairwise
- fuzzy logic
- k means
- initial conditions
- image interpretation
- application layer
- deep learning
- genetic algorithm
- activation function
- belief nets
- automatic initialization
- database
- training process
- recurrent neural networks
- radial basis function
- self organizing maps
- high level
- information systems
- data sets