Performance and Accuracy Tradeoffs for Training Graph Neural Networks on ReRAM-Based Architectures.
Aqeeb Iqbal ArkaBiresh Kumar JoardarJanardhan Rao DoppaPartha Pratim PandeKrishnendu ChakrabartyPublished in: IEEE Trans. Very Large Scale Integr. Syst. (2021)
Keyphrases
- neural network
- training process
- training algorithm
- prediction accuracy
- graph representation
- high accuracy
- random walk
- feedforward neural networks
- pattern recognition
- computational cost
- artificial neural networks
- training set
- graph structure
- genetic algorithm
- neural architectures
- training phase
- spanning tree
- error back propagation
- graph theory
- feed forward
- associative memory
- recurrent neural networks
- directed graph
- graph model
- design decisions
- structured data
- back propagation
- multi layer perceptron
- error rate
- graph theoretic
- fuzzy logic
- neural network structure
- active learning
- computational complexity