SIRIUS-LTG-UiO at SemEval-2018 Task 7: Convolutional Neural Networks with Shortest Dependency Paths for Semantic Relation Extraction and Classification in Scientific Papers.
Farhad NooralahzadehLilja ØvrelidJan Tore LønningPublished in: SemEval@NAACL-HLT (2018)
Keyphrases
- relation extraction
- dependency trees
- semantic features
- scientific papers
- convolutional neural networks
- tree kernels
- information extraction
- domain specific
- automatic extraction
- named entities
- classification accuracy
- image classification
- text classification
- semantic relations
- support vector machine
- feature space
- artificial intelligence
- high level
- machine learning
- broad coverage
- feature selection
- semantic role labeling
- natural language
- word sense disambiguation
- question answering
- feature extraction
- databases
- search engine
- feature set
- supervised learning