A Robust, Quantization-Aware Training Method for Photonic Neural Networks.
A. OikonomouManos KirtasNikos PassalisGeorge Mourgias-AlexandrisMiltiadis Moralis-PegiosNikos PlerosAnastasios TefasPublished in: EANN (2022)
Keyphrases
- neural network
- training process
- computationally efficient
- objective function
- computational complexity
- pairwise
- cost function
- computational cost
- high accuracy
- similarity measure
- quantization error
- highly accurate
- clustering method
- training samples
- significant improvement
- pattern recognition
- detection method
- preprocessing
- training algorithm
- fuzzy neural network
- feed forward neural networks
- neural network structure