MAFAT: Memory-Aware Fusing and Tiling of Neural Networks for Accelerated Edge Inference.
Jackson FarleyAndreas GerstlauerPublished in: IESS (2022)
Keyphrases
- neural network
- auto associative
- pattern recognition
- associative memory
- artificial neural networks
- inference process
- edge detector
- probabilistic inference
- edge detection
- edge information
- back propagation
- bayesian networks
- memory requirements
- main memory
- multi layer
- recurrent neural networks
- fuzzy logic
- rule extraction
- inference engine
- computational power
- feed forward
- random access
- computing power
- memory space
- limited memory
- self organizing maps
- memory size
- genetic algorithm
- data sets
- memory management
- memory usage
- activation function
- edge map
- network architecture
- soft computing
- multilayer perceptron
- neural network model
- feature extraction
- image processing