CorrectNet: Robustness Enhancement of Analog In-Memory Computing for Neural Networks by Error Suppression and Compensation.
Amro EldebikyGrace Li ZhangGeorg BöchererBing LiUlf SchlichtmannPublished in: CoRR (2022)
Keyphrases
- neural network
- auto associative
- artificial neural networks
- error rate
- pattern recognition
- memory requirements
- back propagation
- neural network model
- associative memory
- genetic algorithm
- image enhancement
- analog vlsi
- multi layer
- fuzzy logic
- data sets
- error bounds
- computational efficiency
- multiscale
- fuzzy systems
- activation function
- computing power
- rule extraction
- memory space
- limited memory
- relative error
- fault diagnosis