Generating Authentic Grounded Synthetic Maintenance Work Orders

Large language models (LLMs) are promising for generating synthetic technical data, particularly for industrial maintenance where real datasets are often limited and unbalanced. This study generates synthetic maintenance work orders (MWOs) that are grounded to accurately represent engineering knowle...

Full description

Saved in:
Bibliographic Details
Main Authors: Allison Lau, Jadeyn Feng, Melinda Hodkiewicz, Caitlin Woods, Michael Stewart, Adriano Polpo
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11124200/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849223086678736896
author Allison Lau
Jadeyn Feng
Melinda Hodkiewicz
Caitlin Woods
Michael Stewart
Adriano Polpo
author_facet Allison Lau
Jadeyn Feng
Melinda Hodkiewicz
Caitlin Woods
Michael Stewart
Adriano Polpo
author_sort Allison Lau
collection DOAJ
description Large language models (LLMs) are promising for generating synthetic technical data, particularly for industrial maintenance where real datasets are often limited and unbalanced. This study generates synthetic maintenance work orders (MWOs) that are grounded to accurately represent engineering knowledge and authentic&#x2013;reflecting technician language, jargon, and abbreviations. First, we extracted valid engineering paths from a knowledge graph constructed using the MaintIE gold-annotated industrial MWO dataset. Each path encodes engineering knowledge as a triple. These paths are used to constrain the output of an LLM (<monospace>GPT-4o mini</monospace>) to generate grounded synthetic MWOs using few-shot prompting. The synthetic MWOs are made authentic by incorporating human-like elements, such as contractions, abbreviations, and typos. Evaluation results show that the synthetic data is 86% as natural and 95% as correct as real MWOs. Turing test experiments reveal that subject matter experts could distinguish real from synthetic data only 51% of the time while exhibiting near-zero agreement, indicating random guessing. Statistical hypothesis testing confirms the results from the Turing Test. This research offers a generic approach to extracting legitimate paths from a knowledge graph to ensure that synthetic data generated are grounded in engineering knowledge while reflecting the style and language of the technicians who write them. To enable replication and reuse, code, data and documentation are at <uri>https://github.com/nlp-tlp/LLM-KG-Synthetic-MWO</uri>
format Article
id doaj-art-1e1e4441dabb47e3b5cf792675b85d80
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-1e1e4441dabb47e3b5cf792675b85d802025-08-25T23:12:57ZengIEEEIEEE Access2169-35362025-01-011314588814590410.1109/ACCESS.2025.359875111124200Generating Authentic Grounded Synthetic Maintenance Work OrdersAllison Lau0https://orcid.org/0009-0007-0817-8099Jadeyn Feng1https://orcid.org/0009-0007-5591-153XMelinda Hodkiewicz2https://orcid.org/0000-0002-7336-3932Caitlin Woods3Michael Stewart4Adriano Polpo5https://orcid.org/0000-0002-5959-1808Department of Computer Science and Software Engineering, The University of Western Australia, Perth, WA, AustraliaDepartment of Computer Science and Software Engineering, The University of Western Australia, Perth, WA, AustraliaSchool of Engineering, The University of Western Australia, Perth, WA, AustraliaDepartment of Computer Science and Software Engineering, The University of Western Australia, Perth, WA, AustraliaDepartment of Computer Science and Software Engineering, The University of Western Australia, Perth, WA, AustraliaDepartment of Mathematics and Statistics, The University of Western Australia, Perth, WA, AustraliaLarge language models (LLMs) are promising for generating synthetic technical data, particularly for industrial maintenance where real datasets are often limited and unbalanced. This study generates synthetic maintenance work orders (MWOs) that are grounded to accurately represent engineering knowledge and authentic&#x2013;reflecting technician language, jargon, and abbreviations. First, we extracted valid engineering paths from a knowledge graph constructed using the MaintIE gold-annotated industrial MWO dataset. Each path encodes engineering knowledge as a triple. These paths are used to constrain the output of an LLM (<monospace>GPT-4o mini</monospace>) to generate grounded synthetic MWOs using few-shot prompting. The synthetic MWOs are made authentic by incorporating human-like elements, such as contractions, abbreviations, and typos. Evaluation results show that the synthetic data is 86% as natural and 95% as correct as real MWOs. Turing test experiments reveal that subject matter experts could distinguish real from synthetic data only 51% of the time while exhibiting near-zero agreement, indicating random guessing. Statistical hypothesis testing confirms the results from the Turing Test. This research offers a generic approach to extracting legitimate paths from a knowledge graph to ensure that synthetic data generated are grounded in engineering knowledge while reflecting the style and language of the technicians who write them. To enable replication and reuse, code, data and documentation are at <uri>https://github.com/nlp-tlp/LLM-KG-Synthetic-MWO</uri>https://ieeexplore.ieee.org/document/11124200/Maintenance work orderslarge language modelsGPTknowledge graphsgrounded synthetic datasynthetic data generation
spellingShingle Allison Lau
Jadeyn Feng
Melinda Hodkiewicz
Caitlin Woods
Michael Stewart
Adriano Polpo
Generating Authentic Grounded Synthetic Maintenance Work Orders
IEEE Access
Maintenance work orders
large language models
GPT
knowledge graphs
grounded synthetic data
synthetic data generation
title Generating Authentic Grounded Synthetic Maintenance Work Orders
title_full Generating Authentic Grounded Synthetic Maintenance Work Orders
title_fullStr Generating Authentic Grounded Synthetic Maintenance Work Orders
title_full_unstemmed Generating Authentic Grounded Synthetic Maintenance Work Orders
title_short Generating Authentic Grounded Synthetic Maintenance Work Orders
title_sort generating authentic grounded synthetic maintenance work orders
topic Maintenance work orders
large language models
GPT
knowledge graphs
grounded synthetic data
synthetic data generation
url https://ieeexplore.ieee.org/document/11124200/
work_keys_str_mv AT allisonlau generatingauthenticgroundedsyntheticmaintenanceworkorders
AT jadeynfeng generatingauthenticgroundedsyntheticmaintenanceworkorders
AT melindahodkiewicz generatingauthenticgroundedsyntheticmaintenanceworkorders
AT caitlinwoods generatingauthenticgroundedsyntheticmaintenanceworkorders
AT michaelstewart generatingauthenticgroundedsyntheticmaintenanceworkorders
AT adrianopolpo generatingauthenticgroundedsyntheticmaintenanceworkorders