A Masked Hardware Accelerator for Feed-Forward Neural Networks With Fixed-Point Arithmetic.
Manuel BroschMatthias ProbstMatthias GlaserGeorg SiglPublished in: IEEE Trans. Very Large Scale Integr. Syst. (2024)
Keyphrases
- fixed point
- feed forward neural networks
- floating point
- back propagation
- neural network
- multi layer
- artificial neural networks
- field programmable gate array
- hidden layer
- feed forward
- activation function
- basis functions
- sufficient conditions
- dynamical systems
- hardware implementation
- fuzzy inference system
- learning rate
- training algorithm
- belief propagation
- recurrent neural networks
- floating point unit
- neural nets
- graphical models
- fixed point theorem