The launch of SerenA Academic

As part of our interest in supporting serendipitous connections between researchers, the SerenA project team at Heriot-Watt University have been developing software which uses available information about researchers’ publications and interests. Our approach piggybacks on a growing need for transparent and consistent information systems that allow High Education Institutes (HEIs) to report research activities to external stakeholders, and to offer services to those within and beyond the institutions boundaries. To address this demand, unified data models have been proposed, with a number of software systems emerging to implement these standards. There are currently 19 HEIs that have adopted a system called Pure, which is one such implementation.

We have developed a simple example using the Pure system which has been installed at Heriot-Watt university to gather information about research at Heriot-Watt. “SerenA Academic” takes the information that is made available from this database and builds a semantic representation of the information.

Our demonstration system uses the DBpedia Spotlight extraction tool to derive semantic entities from the abstracts and titles of academic papers. We can find related work by matching these entities in different papers written by different authors. We can also use these connections between their papers to identify connections between researchers in different parts of the university who may share a common interest but do not know about each others’ work. We are aiming to form connections which are seen as interesting, unexpected and valuable.

The data taken from Pure allows us to make connections between academic researchers within Heriot-Watt University. We are exploring ways in which this information can be broadened by providing connections outside the dataset.

First of all, we would like to make connections with researchers who are not represented in the local database. Not all local researchers are represented. In particular, PhD students are not represented, and it would be useful to connect our own PhD students to relevant research papers and academics within the university. Also, researchers outside Heriot-Watt university might wish to connect with researchers inside Heriot-Watt. We are now using the Discover-me-Semantically interface to allow researchers to describe their research expertise, interests and goals in semantic terms. As the form is filled in, an autocompletion feature allows researchers to select easily from the semantic terms used by DBpedia. While researchers select from the English terms, the corresponding semantic represention is stored. Researchers may also paste in an English-language description or abstract of their research, which is analysed by DBpedia Spotlight in the same way as the abstracts in the database. The set of semantic terms supplied or derived from Discover-Me-Semantically is then used to form connections between this researcher and other research papers and academics within Heriot-Watt University.

Second, our semantic representation allows us to pose queries for resources in the larger semantic web. We can make connections with resources which have been described by Wikipedia itself, via their DBpedia entries. So for example if a researcher has expressed interest in Serious_Game, SerenA Academic is able to access DBpedia and find software such as IntelliGym whose genre is also Serious_Game, and returns a brief description of IntelliGym.

We are beginning initial evaluations of SerenA Academic among postgraduate students at Heriot-Watt University and external researchers (members of SerenA!). Researchers use the Discover-Me-Semantically interface and are given a paper portfolio of suggestions for papers to read, academics with related interests, and external resources. We are gathering feedback about the interest, unexpectedness and usefulness of the suggestions made in the portfolio, as well as the types and quantities of information that people need to judge the suggestions.

Leave a Reply

Your email address will not be published. Required fields are marked *



In other news