Graph Convolution for Multimodal Information Extraction from Visually Rich Documents.
Xiaojing LiuFeiyu GaoQiong ZhangHuasha ZhaoPublished in: NAACL-HLT (2) (2019)
Keyphrases
- information extraction
- free text
- text documents
- structured data
- web documents
- information retrieval
- unstructured documents
- textual data
- text mining
- natural language text
- natural language processing
- unstructured text
- graph representation
- semi structured
- xml documents
- precision and recall
- graph model
- document collections
- document classification
- graph theory
- random walk
- directed graph
- relevant documents
- cross document
- graph structure
- metadata
- mutual reinforcement
- information retrieval systems
- machine learning
- weighted graph
- image processing
- document retrieval
- multi document summarization
- named entities
- information extraction systems
- named entity recognition
- multi modal
- audio visual
- vector space model
- web mining
- document clustering
- keywords
- conditional random fields
- semantic information
- multimedia
- question answering
- high level
- text categorization
- graphical models
- similarity measure
- knn
- digital libraries