Accelerating Deep Neural Networks in Processing-in-Memory Platforms: Analog or Digital Approach?
Shaahin AngiziZhezhi HeDayane Alfenas ReisXiaobo Sharon HuWilman TsaiShy Jay LinDeliang FanPublished in: ISVLSI (2019)
Keyphrases
- neural network
- back propagation
- computational power
- real time
- data processing
- circuit design
- memory requirements
- multilayer perceptron
- pattern recognition
- random access
- associative memory
- analog to digital converter
- mixed signal
- processing elements
- memory space
- multi layer
- neural network model
- fuzzy logic
- artificial neural networks
- genetic algorithm
- fuzzy systems
- memory usage
- processing units
- main memory
- limited memory
- digital curves
- processing capabilities
- power consumption
- auto associative
- cmos image sensor