For this thesis topic, you'll investigate methods for using the knowledge embedded in ontologies as reference material for LLMs to increase the accuracy of generated texts.

The main tasks in this project may include:

• Determining appropriate ontologies and LLMs for testing
• Conducting benchmark tests, generating text from LLMs both with and without ontology integration
• Experimenting with different LLM prompts to see how they impact the results
• Assessing the accuracy of the generated texts, highlighting text differences (whether positive, negative, or neutral) that may have resulted from the ontology integration

Depending on the project scope, this topic may be suitable for either a bachelor's thesis or a master's thesis. We'll establish the precise scope and refine the topic in meetings before you begin your work.

Candidates should be have an interest in LLMs, ontologies, knowledge graphs, knowledge engineering, or other related topics.