The current main goals are:
1. Expansion to other disciplines and languages
During the first project phase, we concentrated mostly on German sources from the social sciences domain. InFoLiS II expands this focus on additional scientific and scholarly domains and languages beyond German. In concrete terms, the InfoLink toolkit will be used with various publications and datasets from the fields of social and economic studies and related disciplines in both English and German language. To massively expand the corpus of texts and datasets, we are cooperating with national and international research institutions, repository administrators and publishers.
2. Linked Data based infrastructure
We aim to build a flexible, and long-term sustainable infrastructure to house the algorithms developed in the precursor project for finding links between publications and datasets and embedding them into existing systems.
Based on the paradigms of Linked Data and RESTful web services, all the steps of the InfoLink toolkit will be implemented as self-contained components and workflows will be set into place to automatically update our knowledge base with new data on a regular basis.
We embrace wide-spread usage of the links detected by InfoLink. To work towards wider acceptance, we will develop software to integrate those links into existing research platforms such as publication databases, research data repositories, research infrastructure databases and discovery platforms.
To demonstrate how thusly integrated data improves Information Retrieval, we will build a research prototype that uses the full power of the graph spanned between publications, research data and authors.
3. Improve the reusability of generated links
It is a truth universally acknowledged that the trend to describing, archiving and making available of research data is gaining traction. However, this is a very heterogeneous process.
On the one hand, the granularity (what is the smallest element of research data in need of description?) and the possible, aggregating intermediary steps vary widely. On the other hand, the nature of the relation between publication and dataset varies widely as well.
To formalize this, we will develop an ontology, reusing the DDI-RDF Discovery Vocabulary developed by the Data Documentation Initiative. The resulting improvements of the search process and the reusability of the links in general will also be a focus of the project.