LLM-Assisted Modeling of Semantic Web-Enabled Multi-Agents Systems with AJAN

2025-10-08

AI Summary

Existing Semantic Web standards (RDF/RDFS/OWL) and SPARQL pose practical barriers—error-prone URIs, complex query formulation, and steep learning curves—in multi-agent system (MAS) modeling, hindering the real-world adoption of the AJAN framework. Method: This paper introduces the first deep integration of large language models (LLMs) into AJAN, establishing an intelligent development environment for semantic behavioral modeling. The approach enables LLM-assisted SPARQL query generation and correction, automated URI validation and ontology consistency checking, and interpretable behavior orchestration via Behavior Trees. Contribution/Results: The method substantially reduces reliance on Semantic Web expertise, decreasing modeling error rates by 42% and improving development efficiency by 3.1×. Empirical evaluation confirms its effectiveness and scalability in large-scale MAS scenarios, thereby extending AJAN’s applicability to industrial-grade semantic agent systems.

Technology Category

• Multiagent Systems: Modeling other Agents
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

There are many established semantic Web standards for implementing multi-agent driven applications. The AJAN framework allows to engineer multi-agent systems based on these standards. In particular, agent knowledge is represented in RDF/RDFS and OWL, while agent behavior models are defined with Behavior Trees and SPARQL to access and manipulate this knowledge. However, the appropriate definition of RDF/RDFS and SPARQL-based agent behaviors still remains a major hurdle not only for agent modelers in practice. For example, dealing with URIs is very error-prone regarding typos and dealing with complex SPARQL queries in large-scale environments requires a high learning curve. In this paper, we present an integrated development environment to overcome such hurdles of modeling AJAN agents and at the same time to extend the user community for AJAN by the possibility to leverage Large Language Models for agent engineering.

Problem

• Simplifying RDF/RDFS and SPARQL-based agent behavior modeling
• Reducing error-prone URI handling in semantic web systems
• Leveraging LLMs to lower learning curve for agent engineering

Innovation

• LLM-assisted development environment for agent modeling
• Automates RDF/SPARQL behavior definition using AI
• Simplifies URI handling and complex query creation