An In-Memory-Computing STT-MRAM Macro with Analog ReLU and Pooling Layers for Ultra-High Efficient Neural Network.
Linjun JiangSifan SunJinming GeHe ZhangWang KangPublished in: NVMSA (2023)
Keyphrases
- ultra high
- neural network
- limited memory
- pattern recognition
- artificial neural networks
- fuzzy logic
- auto associative
- databases
- random access memory
- analog circuits
- design considerations
- multi layer
- high resolution
- neural nets
- network architecture
- memory efficient
- multilayer perceptron
- main memory
- data processing
- image classification
- data handling
- genetic algorithm