The Open Knowledge Extraction Challenge for the Semantic Web
PRESENTER:
Amal Zouaq
University of Ottawa
ABSTRACT:
DBpedia represents one of the main hubs of the linked open data cloud and
is increasingly used as a backbone for several knowledge extraction and
enrichment tasks such as semantic annotation and information retrieval. In
fact, the Semantic Web requires the development of techniques that both
exploit current knowledge bases and enrich them using unstructured
sources.
In this talk, I will present our participation to the Open Knowledge
Extraction challenge at the Extended Semantic Web Conference 2016, with
this particular focus in mind. Our two systems, in collaboration with
Polytechnique Montreal, were the winning systems in the
entity typing and linking task (task 2) and in the semantic annotation
task (task 1). In both tasks, DBpedia was used as the
reference knowledge base. I will conclude with a discussion on some of
DBpedia limitations and our current work to overcome some of
these issues.