Service: Enhancing metadata for textual information - improving the quality of search and automatic reasoning

Responsible organisation: Federal Office for Security in Information Technology (Bundesamt für Sicherheit in der Informationstechnik - BSI) (Central-Government)

The solution shall enhance metadata for unstructured, textual information about the topic of IT-Securty and hence rise the quality of search and automatic reasoning. Therefore a text corpus of public available news and background information shall be created and processed such that the pure text content without artifacts and technical accessories is accessible. The resulting texts shall be exploited with a combination of towfold Artificial Intelligence methods. On the one hand traditional methods using manually created taxonomies and ontologies help to find terms of defined meanings in the given text and solve the problem of homonyms. Standards like SKOS (Simple Knowledge Organisation System) guaratny flexibility and reusability. On the other hand machine learning will be applied to determine the local context and support "Named Entity Extraction" (NER). Modern models of Natural Language Processing (NLP) like "Transformer" models will be be applied in various tasks like NER, Question-Answering and classification.

Additional information

Source Open Innovation Regione Lombardia
Web site https://www.bsi.bund.de
Start/end date 2021 -
Still active?

Related cases