The thesis aims to build a practical and theoretical of understanding of knowledge graphs in the age of large language models.
From a practical perspective, we analyse the task of knowledge graph creation using control and experimental models. The experimental models are fully unsupervised or have major components which have been trained in an unsupervised way.