Saved in:
Bibliographic Details
Main Authors: Ai, Kuangshi, Karr Jr, Jonathan A., Jiang, Meng, Chawla, Nitesh V., Wang, Chaoli
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.05524
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918436092772352
author Ai, Kuangshi
Karr Jr, Jonathan A.
Jiang, Meng
Chawla, Nitesh V.
Wang, Chaoli
author_facet Ai, Kuangshi
Karr Jr, Jonathan A.
Jiang, Meng
Chawla, Nitesh V.
Wang, Chaoli
contents We present Knowledge Extraction on OMIn (KEO), a domain-specific knowledge extraction and reasoning framework with large language models (LLMs) in safety-critical contexts. Using the Operations and Maintenance Intelligence (OMIn) dataset, we construct a QA benchmark spanning global sensemaking and actionable maintenance tasks. KEO builds a structured Knowledge Graph (KG) and integrates it into a retrieval-augmented generation (RAG) pipeline, enabling more coherent, dataset-wide reasoning than traditional text-chunk RAG. We evaluate locally deployable LLMs (Gemma-3, Phi-4, Mistral-Nemo) and employ stronger models (GPT-4o, Llama-3.3) as judges. Experiments show that KEO markedly improves global sensemaking by revealing patterns and system-level insights, while text-chunk RAG remains effective for fine-grained procedural tasks requiring localized retrieval. These findings underscore the promise of KG-augmented LLMs for secure, domain-specific QA and their potential in high-stakes reasoning. The code is available at https://github.com/JonathanKarr33/keo.
format Preprint
id arxiv_https___arxiv_org_abs_2510_05524
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle KEO: Knowledge Extraction on OMIn via Knowledge Graphs and RAG for Safety-Critical Aviation Maintenance
Ai, Kuangshi
Karr Jr, Jonathan A.
Jiang, Meng
Chawla, Nitesh V.
Wang, Chaoli
Computation and Language
Information Retrieval
We present Knowledge Extraction on OMIn (KEO), a domain-specific knowledge extraction and reasoning framework with large language models (LLMs) in safety-critical contexts. Using the Operations and Maintenance Intelligence (OMIn) dataset, we construct a QA benchmark spanning global sensemaking and actionable maintenance tasks. KEO builds a structured Knowledge Graph (KG) and integrates it into a retrieval-augmented generation (RAG) pipeline, enabling more coherent, dataset-wide reasoning than traditional text-chunk RAG. We evaluate locally deployable LLMs (Gemma-3, Phi-4, Mistral-Nemo) and employ stronger models (GPT-4o, Llama-3.3) as judges. Experiments show that KEO markedly improves global sensemaking by revealing patterns and system-level insights, while text-chunk RAG remains effective for fine-grained procedural tasks requiring localized retrieval. These findings underscore the promise of KG-augmented LLMs for secure, domain-specific QA and their potential in high-stakes reasoning. The code is available at https://github.com/JonathanKarr33/keo.
title KEO: Knowledge Extraction on OMIn via Knowledge Graphs and RAG for Safety-Critical Aviation Maintenance
topic Computation and Language
Information Retrieval
url https://arxiv.org/abs/2510.05524