Manuela Corte Pause
I'm a grad student in Computer Science at University of Trento and currently working at Ontopic on Virtual Knowledge Graphs
Session
People use natural language for most of their communication needs.
What if people could consult the data in the Open Data Hub using regular natural language.
We will share the experience we gained in the last months, using so-called text2query techniques. We will compare different approaches and benchmark their accuracy in generating a query that matches as well as possible the intent of the user.
More specifically, we will compare the following approaches:
1) directly the underlying data source in SQL
2) the virtual knowledge graph in SPARQL
3) the virtual knowledge graph in SQL.
Our research in this area is an ongoing effort for the ERDF project "CRIMA", a joint project of Ontopic, unibz and Eurac.