An Overview of Knowledge Extraction Projects in the NLP group at Lund University

In this paper, I describe systems and prototypes we created in the natural language processing group at Lund to extract structured knowledge from text. Starting from syntactic and semantic parsingcomponents,wedevelopedapplicationsthatcanhandlelargecorpora,typicallycomplete Wikipedia versions consisting of millions of documents and process text to identify entities and the relations between them. I describe the overall goals of our projects, the data structure we designed to handle the documents, as well as three applications to extract knowledge from text