Doc-GCN: Heterogeneous Graph Convolutional Networks for Document Layout Analysis.
Siwen LuoYihao DingSiqu LongJosiah PoonSoyeon Caren HanPublished in: COLING (2022)
Keyphrases
- heterogeneous networks
- average degree
- information retrieval
- social networks
- directed graph
- document collections
- document images
- structured data
- fully connected
- graph structures
- graph representation
- graph layout
- graph theory
- protein interaction networks
- document representation
- betweenness centrality
- document clustering
- complex networks
- information retrieval systems
- random walk
- web documents
- search engine
- connected components
- keywords
- social graphs
- degree distribution
- community discovery
- edge weights
- weighted graph
- document retrieval
- text documents
- graph structure
- network analysis
- power law
- small world
- path length
- deep learning
- user queries
- retrieval systems
- information extraction
- multiple types
- bipartite graph
- xml documents
- graph matching
- community structure
- biological networks