EXPAND

In a data-driven society where decision-making depends on advanced data analysis, valuable information is often scattered across multiple sources. This raises the challenge of providing an integrated view of multiple data sources to allow for intuitive and efficient querying by a stakeholder. The EXPAND project aims to facilitate data access and integration through the Ontology-Based Data Access (OBDA) technology. In a nutshell, OBDA provides a principled way of integrating multiple data sources by adding an ontological layer on top of them. The foundations of OBDA have been the subject of intense research over the last 15 years, and have already allowed for successful industrial applications. However, despite its promises, the current OBDA framework still faces some sensible challenges. Indeed, to reap the full potential of the approach, three key ingredients are necessary: first, the user should be able to express their information need, and this requires expressive query languages; second, the user should be able to understand the reason why a given answer is provided, and this requires both explanation facilities and compact representations of diverse answers ; third, the system should provide answers in a timely manner, and this requires efficient querying algorithms that scale in the presence of large datasets. The state of the art in OBDA is still severely limited regarding these three challenges. The EXPAND project will tackle these challenges by pursuing three goals. The first goal is to extend OBDA to support query classes featuring aggregation, navigation and default negation. The second goal is to provide enough context to a user to help them to accept or discard answers provided by the system, in particular by leveraging techniques from neighbouring fields. The third goal is to develop data-aware optimization techniques to make an expressive OBDA framework applicable in practice. Finally, we will demonstrate our techniques on three use cases in agriculture and agronomy, building upon ongoing collaborations with INRAe and DFKI.