Knowledge graphs are flexible but structured data types to store relations between entities and
concepts. Auto-generation of these graphs is an active research field that enables representing
large amounts of text data efficiently and solving other NLU problems such as text
summarization, and question answering with the help of graph algorithms and graph neural
networks. This project is structured on the question of how knowledge graphs can be generated
from unstructured text input and what NLP components are needed to be utilized. Furthermore,
representing entities and relations in a graph embedding space to be able to generate new
relations was another challenge. NLP researchers find answers to these questions under Open
Information Extraction, Entity Linking, Knowledge Graph Embeddings and Knowledge Graph
Completion study fields.
