TC-GNN: Accelerating Sparse Graph Neural Network Computation Via Dense Tensor Core on GPUs.
Yuke WangBoyuan FengYufei DingPublished in: CoRR (2021)
Keyphrases
- neural network
- densely connected
- high order
- tensor decomposition
- graph theory
- higher order
- graph representation
- graph structure
- weighted graph
- motion field estimation
- gaussian graphical models
- tensor factorization
- directed acyclic graph
- artificial neural networks
- dense stereo
- neural network model
- directed graph
- graph mining
- graph model
- quasi cliques
- parallel computation
- dense optical flow
- neural network is trained
- high dimensional
- general purpose
- genetic algorithm
- pattern recognition
- auxiliary information
- bipartite graph
- dimensionality reduction
- random walk
- recurrent neural networks