Graph Policy Network for Transferable Active Learning on Graphs.
Shengding HuZheng XiongMeng QuXingdi YuanMarc-Alexandre CôtéZhiyuan LiuJian TangPublished in: NeurIPS (2020)
Keyphrases
- active learning
- random graphs
- spanning tree
- graph theory
- small world
- highly connected
- graph matching
- finding the shortest path
- graph representation
- fully connected
- dynamic networks
- graph theoretic
- graph structure
- directed graph
- degree distribution
- community discovery
- graph databases
- edge weights
- weighted graph
- graph model
- average degree
- graph mining
- graph clustering
- graph theoretical
- labeled graphs
- social graphs
- graph partitioning
- dense subgraphs
- graph structures
- bipartite graph
- graph properties
- graph construction
- undirected graph
- real world social networks
- graph data
- strongly connected
- graph representations
- link prediction
- subgraph isomorphism
- community structure
- graph isomorphism
- graph search
- real world networks
- connected subgraphs
- complex networks
- random walk
- structural pattern recognition
- power law
- graph layout
- densely connected
- social networks
- graph classification
- dynamic graph
- series parallel
- adjacency matrix
- semi supervised
- connected components
- optimal policy
- graph mining algorithms
- planar graphs
- directed edges
- machine learning
- massive graphs
- training set
- semi supervised learning
- structured data
- reachability queries
- directed acyclic graph
- path length
- minimum spanning tree
- protein interaction networks
- maximal cliques
- clustering coefficient
- network properties