Probabilistic Language Models for the Creation of Enterprise Knowledge Graphs

Abstract

This master thesis explores the use of probabilistic (large) language models (LLMs) in ontology and knowledge graph construction for enterprises. It thereby addresses the challenges associated with conventional approaches, such as the high level of manual effort and the need for specialist expertise. The study adopts an inductive design science research strategy and applies a qualitative research methodology. For data collection, interviews with knowledge engineering experts and documents analysis are conducted.