Write-Energy Relaxation of MTJ-Based Quantized Neural-Network Hardware.
Ken AsanoMasanori NatsuiTakahiro HanyuPublished in: ISMVL (2023)
Keyphrases
- neural network
- low cost
- artificial neural networks
- pattern recognition
- hardware and software
- neural network model
- neural nets
- back propagation
- energy consumption
- fault diagnosis
- probabilistic relaxation
- energy minimization
- prediction model
- learning vector quantization
- self organizing maps
- real time
- genetic algorithm
- training algorithm
- energy efficiency
- neural network is trained
- hardware architecture
- convex relaxation
- computing systems
- multi layer
- massively parallel
- energy saving
- feed forward neural networks
- personal computer
- associative memory
- vlsi implementation
- feed forward
- objective function