Faculty of Computer Science

Research Group Theoretical Computer Science


Oberseminar: Heterogene formale Methoden


Date: 2022, May 10
Time: 10:30 a. m.
Place: online
Author: Razzaghian, Negar
Title: Translation of Natural Language Competency Questions into SPARQL-OWL Queries Using Neural Language Models

Abstract:

Competency questions are a set of questions in natural language about ontologies, that determine the main scope of a specific ontology. A machine translation of questions simplifies the formulations of them in SPARQL-OWL queries. There are many embedding approaches, that are the main input for the machine encoder. The Word embedding approach is one of the most common approaches, that follows an embedding of vocabularies in a vector representation and feed that to the neural language machine. Word embedding approach applies skip-gram algorithm to sample and generate the training data. This presentation explains about word embedding approach and the skip-gram algorithm.


Back to the Oberseminar web page
Webmaster