Monolithic Silicon Photonic Architecture for Training Deep Neural Networks with Direct Feedback Alignment.
Matthew J. FilipovichZhimu GuoMohammed Al-QadasiBicky A. MarquezHugh D. MorisonVolker J. SorgerPaul R. PrucnalSudip ShekharBhavin J. ShastriPublished in: CoRR (2021)
Keyphrases
- neural network
- training algorithm
- training process
- multi layer
- feed forward neural networks
- network architecture
- error back propagation
- neural network structure
- neural network training
- backpropagation algorithm
- management system
- echo state networks
- feedforward neural networks
- multi layer perceptron
- back propagation
- associative memory
- multilayer neural network
- high speed
- training sessions
- deep architectures
- feedforward artificial neural networks
- genetic algorithm
- pattern recognition
- recurrent neural networks
- training set
- artificial neural networks
- multilayer perceptron
- supervised learning
- recurrent networks
- relevance feedback
- test set
- software architecture
- training phase
- neural nets
- radial basis function network
- dynamic time warping
- fault diagnosis
- x ray
- low cost
- image alignment
- fuzzy logic
- high density