Understanding the Representation Power of Graph Neural Networks in Learning Graph Topology.
Nima DehmamyAlbert-László BarabásiRose YuPublished in: NeurIPS (2019)
Keyphrases
- neural network
- graph representation
- fully connected
- learning algorithm
- graph structure
- connected components
- supervised learning
- graphical representation
- graph model
- graph mining
- learning systems
- image representation
- connectionist systems
- graph structures
- active learning
- learning process
- graph matching
- incremental learning
- graph theoretic
- learning tasks
- error back propagation
- topological information
- bayesian networks
- subject matter
- deeper understanding
- neural nets
- pattern recognition
- learning rules
- graph partitioning
- spanning tree
- online learning
- knowledge acquisition
- directed acyclic graph