Layer-Wise Interpretation of Deep Neural Networks using Identity Initialization.
Shohei KubotaHideaki HayashiTomohiro HayaseSeiichi UchidaPublished in: ICASSP (2021)
Keyphrases
- neural network
- multi layer
- multiple layers
- single layer
- image interpretation
- pattern recognition
- identity management
- application layer
- neural nets
- artificial neural networks
- semantic interpretation
- fuzzy logic
- feed forward
- genetic algorithm
- feed forward neural networks
- activation function
- hidden layer
- multilayer perceptron
- back propagation
- k means
- high level
- metadata
- deep learning
- database
- social context
- rule extraction
- network architecture
- training process
- neural network model
- pairwise
- face recognition