GDR-HGNN: A Heterogeneous Graph Neural Networks Accelerator Frontend with Graph Decoupling and Recoupling.
Runzhen XueMingyu YanDengke HanYihan TengZhimin TangXiaochun YeDongrui FanPublished in: CoRR (2024)
Keyphrases
- neural network
- random walk
- graph representation
- pattern recognition
- structured data
- directed graph
- graph structure
- graph theory
- stable set
- databases
- graph model
- connected components
- data sets
- graph partitioning
- weighted graph
- graph matching
- database
- bipartite graph
- neural nets
- directed acyclic graph
- back propagation
- fuzzy logic
- probabilistic model
- graph theoretic
- graph construction
- adjacency matrix