FlowGNN: A Dataflow Architecture for Universal Graph Neural Network Inference via Multi-Queue Streaming.
Rishov SarkarStefan Abi-KaramYuqi HeLakshmi SathideviCong HaoPublished in: CoRR (2022)
Keyphrases
- neural network
- real time
- inference engine
- network architecture
- data flow
- bayesian networks
- neural architecture
- design methodology
- back propagation
- graph structure
- multi layer
- neural network model
- database machine
- artificial neural networks
- management system
- directed acyclic graph
- graph transformation
- graph theory
- data streams
- dynamic reconfiguration
- probabilistic inference
- recurrent neural networks
- video streaming
- graph representation
- directed graph
- steady state
- fault diagnosis
- associative memory
- bayesian inference
- graph matching
- software architecture
- graph databases
- self organizing maps
- markov chain
- graphical models
- fuzzy logic
- expert systems
- pattern recognition
- genetic algorithm
- neural network is trained