Abstract—Predicate-argument structures are best known as means to represent shallow semantics behind natural language sentences by employing semantic role labeling (SRL) technique. The latter serves as foundation for complex tasks like question answering, text summarization, plagiarism detection and others. In this paper we show how SRL and semantic web technology can be used to build a knowledge graph from open-domain natural language texts with the main goal of enabling semantically-flavored information retrieval on top of the resulting knowledge base. In particular, we propose a domain-agnostic ontology schema capable of capturing event-oriented knowledge and a modification of breadth-first search graph traversal algorithm for serving users information needs. Finally, we evaluate behavior of the whole framework by annotating part of WikiQA dataset and use the constructed knowledge graph to judge information retrieval effectiveness which shows promising results.
Index Terms—Semantic search, knowledge graphs, RDF, semantic role labeling.
Tomas Vileiniskis and Rita Butkiene are with Kaunas University of Technology, Department of Information Systems, Studentu st. 50, Kaunas, Lithuania (e-mail: tomas.vileiniskis@ktu.edu; rita.butkiene@ktu.edu).
[PDF]
Cite: Tomas Vileiniskis and Rita Butkiene, "Leveraging Predicate-Argument Structures for Knowledge Extraction and Searchable Representation Using RDF," International Journal of Knowledge Engineering vol. 6, no. 1, pp. 30-34, 2020.
Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).