LAMA-WeST Lab

Artificial Intelligence
Natural language processing
Semantic web

Extraction of axioms and logical rule from Wikipeda definitions

Student: Lara Haidar-Ahmad

Supervisor: Amal Zouaq

Co-supervisor(s): Michel Gagnon

The Semantic Web relies on the creation of rich knowledge bases which links data on the Web. In that matter, DBpedia started as a community effort and is considered today as the central interlinking hub for the emerging Web of data. However, DBpedia relies on a lighweight ontology and deals with some substantial limitations and lacks some important information that could be found in the text and the unstructured data of Wikipedia. Furthermore, the DBpedia ontology contains mainly taxonomical links and data about the instances, and lacks class definitions. The objective of this work is to enrich DBpedia with class definitions and taxonomical links using text in natural language. For this purpose, we rely on a pattern-based approach that transforms textual definitions from Wikipedia into RDF graphs, which are processed to query syntactical pattern occurrences using SPARQL. This work resulted in the creation of AXIOpedia, a rich knowledge base containing complex axioms defining classes and rdf:type relations relating instances with these classes.

For more information, click here

Publications

2016
July

Automatic Extraction of Axioms from Wikipedia Using SPARQL

Lara Haidar-Ahmad, Amal Zouaq, Michel Gagnon

Lara Haidar-Ahmad, Amal Zouaq, Michel Gagnon

2016
May

Entity Typing and Linking using SPARQL Patterns and DBpedia

Lara Haidar-Ahmad, Ludovic Font, Amal Zouaq, M. Gagnon

ESWC-16 Open Knowledge Extraction Challenge, Heraklion, Greece