Unlocking the Power of LLM-Based Question Answering Systems: Enhancing Reasoning, Insight, and Automation with Knowledge Graphs

In today’s data-driven business landscape, Knowledge Graphs can be effectively layered on top of relational databases and ontologies, a powerful combination for transforming how businesses tackle complex queries and decision-making processes. In this paper, we present a series of experiments that demonstrate the opportunities and advantages of blending knowledge graphs with Large Language Models (LLMs) through a practical use case. Our experimental results provide insights into the reasoning capabilities of LLMs when utilizing Knowledge Graph-Prompting. Furthermore, we observed the significance of maintaining uniformity in the language employed during knowledge graph construction to ensure precise responses from LLMs when querying the knowledge graph. This consistency also resonates in the embedding space of the model, where elements like relationship types are reflected in the resulting embeddings.