Multi-Agent System for Conversational Knowledge Graph Interaction

This thesis focuses on developing a multi-agent system that enables users to interact with a knowledge graph using natural language queries. The system will consist of specialized agents that communicate through semantic web interfaces, each performing distinct roles such as data retrieval, result analysis, and user-friendly visualization. The goal is to create an autonomous, modular architecture where agents collaboratively process and present relevant information.

A foundational version of this multi-agent system already exists, providing a strong starting point for further development. The thesis will extend its capabilities by refining agent collaboration, improving response accuracy, and enhancing usability for real-world applications. Potential research directions include optimizing agent communication, integrating advanced reasoning techniques, and improving the system’s ability to handle complex queries efficiently.

By building on existing work, this research aims to push the boundaries of conversational AI for knowledge graphs, making database interaction more intuitive and accessible for users.

This thesis is conducted in cooperation with Adorsys. If you are interested in this topic, please contact Mark Kram (mark.kram@adorsys.com).