From Monday 8th of October to Wednesday 17th of October, NLP Town organized a Hackathon on Natural Language Processing with a focus on legal and administrative NLP (Natural Language Processing). The hackathon was hosted by Informatie Vlaanderen and sponsored by technology leader SAP. Startups, students, professionals and hobby NLP'ers had the opportunity to develop prototype NLP applications that can make the government smarter and more efficient.
The hackathon took the form of a sprint, in which teams worked on a targeted project for 10 days. On Monday 8th of October, the first day of the sprint, Flemish government services presented 20 use cases where NLP (language technology) could possibly help them. After this introduction, the teams could start designing and implementing their applications. Sopra Steria took on the challenge of using NLP to categorize the information from the recent elections, which is available in the form of unstructured documents. The objective was to solve who takes which mandate in which community for which party by extracting entities from a large set of different document types. The solution was compliant with the OSLO model, which is the preferred solution by the Vlaamse Overheid.
The team presented their POC (Proof of Concept) in an explanatory video to the government departments participating in the hackaton. The jury selected the most promising application on three different categories. Sopra Steria was recognized due to their Vlaamse Overheid oriented approach as one of the most promising project.
We would like to congratulate the team representing Sopra Steria for their hard work and commitment to the event.