CANN: Curable Approximations for High-Performance Deep Neural Network Accelerators.
Muhammad Abdullah HanifFaiq KhalidMuhammad ShafiquePublished in: DAC (2019)
Keyphrases
- neural network
- back propagation
- artificial neural networks
- genetic algorithm
- pattern recognition
- graphics processing units
- neural network model
- neural nets
- efficient implementation
- deep learning
- multi layer perceptron
- computing systems
- training algorithm
- network model
- image processing
- radial basis function
- self organizing maps
- image reconstruction from projections
- neural network is trained
- high reliability
- computing platform
- general purpose
- efficient computation
- fault diagnosis
- parallel processing
- cost effective
- hidden layer
- feed forward
- bp neural network
- prediction model
- real time