IDLab (Internet & Data Lab, University of Ghent)
milieu (Law & Policy Consulting, Belgium)
The project develops the RCD (Regulatory Concept Dictionary) application for DG FISMA, the department of the European Commission that develops and carries out the latter’s policies on financial services.
The application provides support for a greater standardisation of EU-level supervisory reporting requirements, by automatically creating a glossary of concepts defined in the body of legal texts within the domain of DG FISMA and setting up a dictionary of reporting obligations (specifying who reports what to whom, at what time, etc.) contained in these texts.
The application allows for automatically collecting English documents (web pages and files) from websites and detecting the documents relevant to DG FISMA. The application can be used as a search engine restricted to the domain of DG FISMA.
Various interfaces show concepts, definitions, reporting obligations, and their (highlighted) occurrences and allow for searching concepts and reporting obligations (based on criteria like “reporting entity”). The application makes use of natural language processing and machine learning.
As the reporting obligation extraction task is inherently complex, further improvements are needed. The user of the application can provide various types of feedback, for instance by changing a definition’s text span. A developer can retrain and improve components of the application based on users’ feedback.
The final report of the project is published here.