Ontologies comprising of concepts, their attributes, and relationships are used in many knowledge based AI systems. While there have been efforts towards populating domain specific ontologies, we examine the role of document structure in learning ontological relationships between concepts in any document corpus. Inspired by ideas from hypernym discovery and explainability, our method performs about 15 points more accurate than a stand-alone R-GCN model for this task.
@article{arxiv.2104.12950,
title = {Document Structure aware Relational Graph Convolutional Networks for Ontology Population},
author = {Abhay M Shalghar and Ayush Kumar and Balaji Ganesan and Aswin Kannan and Akshay Parekh and Shobha G},
journal= {arXiv preprint arXiv:2104.12950},
year = {2022}
}
Comments
8 pages single column, 5 figures. DLG4NLP Workshop at ICLR 2022