Dataflow-based Joint Quantization of Weights and Activations for Deep Neural Networks.
Xue GengJie FuBin ZhaoJie LinMohamed M. Sabry AlyChristopher J. PalVijay ChandrasekharPublished in: CoRR (2019)
Keyphrases
- neural network
- pattern recognition
- artificial neural networks
- linear combination
- connection weights
- weight update
- database machine
- back propagation
- neural nets
- genetic algorithm
- weighted sum
- data flow
- multilayer perceptron
- quantization error
- neural network model
- hidden neurons
- weighting scheme
- parallel computing
- relative importance
- multi layer
- recurrent neural networks
- design methodology
- activation function
- fuzzy logic
- associative memory
- high dimensional
- learning algorithm
- sufficient conditions