This research introduces a hybrid approach combining neural networks with symbolic reasoning to address a critical LLM limitation: hallucinations.

• Integrates OWL ontologies with symbolic reasoners to verify factual consistency
• Creates a technical pipeline that filters LLM outputs through formal knowledge structures
• Significantly improves reliability in domains requiring factual accuracy
• Demonstrates how engineering disciplines can combine deep learning with traditional logical systems

For engineering applications, this approach offers a practical method to implement LLMs in high-stakes environments where factual errors could have serious consequences.

Enhancing Large Language Models through Neuro-Symbolic Integration and Ontological Reasoning