Memory Requirement Reduction of Deep Neural Networks Using Low-bit Quantization of Parameters.
Niccoló NicodemoGaurav NaithaniKonstantinos DrossosTuomas VirtanenRoberto SalettiPublished in: CoRR (2019)
Keyphrases
- memory requirements
- neural network
- memory space
- computational complexity
- memory usage
- pattern recognition
- computational speed
- storage requirements
- highly non linear
- parameter values
- computational power
- sensitivity analysis
- fuzzy logic
- artificial neural networks
- adaptive quantization
- network latency
- color quantization
- neural nets
- fuzzy inference system
- parameter estimation
- sufficient conditions
- genetic algorithm