3D-ReG: A 3D ReRAM-based Heterogeneous Architecture for Training Deep Neural Networks.
Bing LiJanardhan Rao DoppaPartha Pratim PandeKrishnendu ChakrabartyJoe X. QiuHai (Helen) LiPublished in: ACM J. Emerg. Technol. Comput. Syst. (2020)
Keyphrases
- neural network
- training process
- neural network structure
- network architecture
- training algorithm
- multi layer
- feed forward neural networks
- feedforward neural networks
- real time
- multilayer neural network
- management system
- neural architectures
- multi layer perceptron
- pattern recognition
- backpropagation algorithm
- neural network training
- deep architectures
- genetic algorithm
- loosely coupled
- neural models
- training samples
- back propagation
- software architecture
- neural network model
- artificial neural networks
- error back propagation
- radial basis function network
- learning capabilities
- feedforward artificial neural networks
- training phase
- supervised learning
- training examples
- online learning
- data sets
- real world
- training data
- feature space
- training set
- fuzzy logic
- recurrent networks
- training patterns
- test set
- fault diagnosis
- feed forward
- multilayer perceptron
- fuzzy neural network