Using Graph Neural Networks to Reconstruct Ancient Documents.
Cecilia OstertagMarie Beurton-AimarPublished in: ICPR Workshops (7) (2020)
Keyphrases
- neural network
- information retrieval
- mutual reinforcement
- document collections
- graph representation
- relevant documents
- text documents
- artificial neural networks
- pattern recognition
- document retrieval
- information retrieval systems
- random walk
- document clustering
- neural network model
- weighted graph
- graph structure
- graph theory
- metadata
- web documents
- database
- directed graph
- xml documents
- structured data
- document classification
- fuzzy logic
- connected components
- electronic documents
- directed acyclic graph
- graph model
- document analysis
- cultural heritage
- vector space
- recurrent neural networks
- feed forward
- retrieval systems
- semantic information
- three dimensional
- learning algorithm
- graph matching
- vector space model
- graph mining
- back propagation
- wordnet
- multimedia documents
- keywords
- similarity scores
- citation graph