C-DNN: A 24.5-85.8TOPS/W Complementary-Deep-Neural-Network Processor with Heterogeneous CNN/SNN Core Architecture and Forward-Gradient-Based Sparsity Generation.
Sangyeob KimSoyeon KimSeongyon HongSangjin KimDonghyeon HanHoi-Jun YooPublished in: ISSCC (2023)
Keyphrases
- neural network
- training algorithm
- network architecture
- loosely coupled
- nearest neighbor
- training process
- back propagation
- parallel architecture
- cellular neural networks
- industry standard
- neural architecture
- multi layer
- management system
- instruction set
- high dimensional
- artificial neural networks
- computation intensive
- real time
- multilayer perceptron
- parallel processing
- software architecture
- multi processor
- neural nets
- memory hierarchy
- genetic algorithm
- computer architecture
- processing elements
- level parallelism
- fuzzy logic
- convolutional neural network
- fuzzy neural network
- associative memory
- operating system
- high speed