The "VIVO Research Information Platform" project aims to present research metadata within the Berlin University Alliance (BUA) in one platform and to make it searchable. This is made possible by ontologies that semantically link the individual contents of the platform. Rolf Guescini, IT developer on the project, is leading the ontology development for the platform. In this interview, he provides insight into the functionality, significance, challenges and opportunities.

 

Rolf, how would you explain the term "ontology" as simple as possible?

The term "ontology" is defined differently in various disciplines. In computer science we use it to describe a certain domain of knowledge. An Ontology contains terms that represent a concept in the world. This can be concrete or abstract. At the same time, the ontology includes the relationships between different concepts. In simple terms, an ontology includes concepts and relationships within a domain of knowledge. It is like the glue between different information objects.

 

How do ontologies work?

It is an identifier-based technology that assigns an identifier to each concept. For example, an identifier is assigned to a horse, so we can talk to each other about a horse without any doubt. The idea behind this is that we can talk unambiguously about a concept anywhere in the world and research on it can be mapped collectively. In reality, of course, the situation is much more complex.

 

What role do ontologies play for the "VIVO Research Information Platform" for the Berlin University Alliance?

The "FIP with VIVO" is a semantic platform based on ontologies. Using the open source software VIVO, we can provide end users with a framework to interact with these ontologies in a simple environment.

For this purpose, all information objects, so-called entities, for example people, publications or organizational elements, are represented in ontologies. The ontologies in turn form networks, called graphs, which can be queried using the query language "SPARQL". The connections of the graphs allow us to discover and visualize new, even surprising connections between entities in the platform.

 

Which ontologies are used in the "VIVO Research Information Platform"?

Within the VIVO software there are several ontologies representing different aspects - researchers themselves, their research, publications and organizational structure.

We use the VIVO ontology and extend it for the European area. We also had to develop an ontology that represents the specific organizational structure of the four partner organizations of the BUA.

At the same time, we can rely on the work of various projects to categorize publications and, hopefully in the future, also projects and research data. For disciplinary classification of research fields we use the ontology of B2FIND EUDAT. For interdisciplinary research field classifications, we draw on the ontology of the Kerndatensatz Forschung in Germany.

 

What challenges are you facing during ontology development?

The VIVO software was developed for the American academic space. Accordingly, the VIVO ontology describes organizational elements and roles for this language area. On the one hand, different terms are used at European institutions, and on the other hand, names are used that are similar to American terms but have a different meaning than their English-language equivalents. Therefore, the existing ontology had to be extended to describe European structures with the correct semantics.

The challenge in developing the BUA ontology is quite similar: the same terms, such as for a department, sometimes have different meanings among BUA's partner organizations. Accordingly, we need to harmonize the ontologies so that we can query and compile interdisciplinary and cross-institutional results in the platform.

 

Is the BUA ontology extension useful for other projects in the VIVO community?

Each project, of course, has local attributes that are difficult to translate into a larger, generic context. However, since the VIVO community builds its development based on "upper" or generic ontologies that any project can inherit from, we can assume that our extensions will be usable for similar projects in the same language space.

 

What is the status now and what are the next steps?

We have currently listed people, their publications and we are linking people to the organizational structures.

In the future, we want to link people to projects and projects to their respective research outputs. We also want to link these structures to other projects within the BUA, such as the Shared Services Catalogue.

 

What can we accomplish by doing this?

We will achieve the greatest value for all collaborative partners if we link the different ontologies within the BUA and thus bring them into a larger context.

By this it will be possible to link and present research in a field using common categorizations and common concepts - across disciplines and institutional boundaries.

When we work with such linked and broad ontologies, we can use the search function to make connections and find content that otherwise would not be findable. In this way, we hope to connect, visualize, and ultimately support interdisciplinary research.