Intent-Driven Smart Manufacturing Integrating Knowledge Graphs and Large Language Models

2026-02-16 · Source: cs.AI updates on arXiv.org · Field: Manufacturing & Industrial — Smart Manufacturing & Industry 4.0, Artificial Intelligence & Machine Learning, Manufacturing Operations & Management · Depth: Advanced, extended

Summary

A new framework 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. Researchers fine-tuned Mistral-7B-Instruct-V02 on a custom dataset of 2,580 domain-specific samples, achieving 89.33% exact match accuracy and 97.27% overall accuracy in translating natural language intents into structured JSON requirement models. These models are then semantically mapped to a Neo4j-based knowledge graph, grounded in the ISA-95 standard, to ensure operational alignment with manufacturing processes, resources, and constraints. The system significantly outperforms zero-shot and few-shot baselines, demonstrating its potential for scalable, explainable, and adaptive human-machine collaboration in smart manufacturing, with an average inference time of 54ms per sample.

Key takeaway

For AI Scientists developing smart manufacturing solutions, this framework offers a robust method to bridge natural language user intents with machine-executable actions. You should consider fine-tuning lightweight LLMs like Mistral-7B-Instruct-V02 on domain-specific datasets and integrating them with an ISA-95 compliant Neo4j knowledge graph to achieve high accuracy and real-time performance in MaaS environments. This approach enhances interpretability and operational flexibility, crucial for dynamic industrial settings.

Key insights

Integrating fine-tuned LLMs with KGs translates natural language manufacturing intents into machine-executable actions with high accuracy.

Principles

• Instruction tuning improves LLM performance for domain-specific structured output.
• KGs provide structured context for operationalizing LLM-generated intents.

Method

A multi-stage pipeline translates natural language intents into JSON requirement models via a fine-tuned LLM, then maps these to a Neo4j-based KG using an ISA-95 aligned ontology for operational execution.

In practice

• Fine-tune Mistral-7B-Instruct-V02 for structured JSON output.
• Use 4-bit quantization and LoRA for efficient LLM training.
• Ground manufacturing KGs in ISA-95 for operational alignment.

Topics

• Smart Manufacturing
• Intent-Driven Systems
• Large Language Models
• Knowledge Graphs
• LLM Fine-tuning

Best for: NLP Engineer, AI Scientist, AI Engineer, Machine Learning Engineer, Research Scientist

Related on AIssential

• Using Large Language Models and Knowledge Graphs to Improve the Interpretability of Machine Learning Models in Manufacturing — Combining LLMs and KGs can generate user-friendly, context-aware explanations for ML model predictions.
• FineREX: Fine-Tuned NER-RE for Human Smuggling Knowledge Graphs — Domain-specific fine-tuning of LLMs substantially improves knowledge graph quality and efficiency for complex legal text.
• : Structuring Your Natural Language SOPs into Tailored Ambiguity-Resolved Code Templates — A multi-agent LLM framework effectively transforms ambiguous SOPs into structured, executable code with high accuracy.
• SING: Synthetic Intention Graph for Scalable Active Tool Discovery in LLM Agents — Intention-aware graph structures enable accurate, context-efficient tool discovery for LLM agents in large ecosystems.
• Providing Greater Access to LLMs with Brandon Duderstadt, Co-Founder and CEO of Nomic AI — Fine-tuning and prompt engineering are crucial for optimizing LLM performance and accessibility.
• Benchmarking Local LLMs for Natural-Language-to-SQL Querying in Biopharmaceutical Manufacturing: An Empirical Benchmark on Consumer-Grade Hardware — Code-tuned general-purpose LLMs outperform domain-specific medical LLMs for T-SQL generation on manufacturing schemas.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.