Intent-Driven Smart Manufacturing Integrating Knowledge Graphs and Large Language Models
Takoua Jradi, John Violos, Dimitrios Spatharakis, Lydia Mavraidi, Ioannis Dimolitsas, Aris Leivadeas, Symeon Papavassiliou
TL;DR
This work tackles the challenge of converting high-level human intents into machine-executable actions within Manufacturing-as-a-Service by coupling an instruction-tuned LLM with a domain ISA-95–aligned knowledge graph. A three-stage pipeline translates natural-language intents into structured JSON requirement models, aligns them with a manufacturing ontology, and updates a Neo4j KG to enable context-aware reasoning and execution. Fine-tuning the Mistral-7B-Instruct-V02 on 2,580 domain-specific samples yields strong performance, achieving 89.33% exact-match accuracy and 97.27% overall accuracy compared to baselines. The framework supports explainable, adaptive human-machine interaction with scalable KG updates and sets the stage for retrieval-augmented generation and online adaptation in MaaS ecosystems.
Abstract
The increasing complexity of smart manufacturing environments demands interfaces that can translate high-level human intents into machine-executable actions. This paper presents a unified framework that integrates instruction-tuned Large Language Models (LLMs) with ontology-aligned Knowledge Graphs (KGs) to enable intent-driven interaction in Manufacturing-as-a-Service (MaaS) ecosystems. We fine-tune Mistral-7B-Instruct-V02 on a domain-specific dataset, enabling the translation of natural language intents into structured JSON requirement models. These models are semantically mapped to a Neo4j-based knowledge graph grounded in the ISA-95 standard, ensuring operational alignment with manufacturing processes, resources, and constraints. Our experimental results demonstrate significant performance gains over zero-shot and 3-shots baselines, achieving 89.33\% exact match accuracy and 97.27\% overall accuracy. This work lays the foundation for scalable, explainable, and adaptive human-machine