Detecting and Fixing Inconsistencies in Large Knowledge Graphs*
- 8 July 2025
- journal article
- Published by SAGE Publications in The European Journal on Artificial Intelligence
- Vol. 39 (2) , 251-269
Abstract
Lately, the availability of massive amounts of data necessitates the adoption of modern representation techniques, such as knowledge graphs (KGs). KGs are typically constructed via automated procedures and by utilizing heterogeneous data sources. This nevertheless hinders the quality of these large resulting KGs, as they might contain contradictions, that is a set of assertions that conflict with some axioms often set by human experts. In turn, classical description logics reasoners cannot be applied as no useful inference results can be generated in the face of inconsistencies. Meanwhile, classical reasoners can be used in order to retrieve the inconsistency explanations, but as the KG size grows larger the required time for this task increases significantly. To address the problem of reasoning with large and inconsistent KGs, we propose an open-source system that detects and fixes inconsistencies by splitting the KG into modules and then processing them in parallel to speed up the process. An empirical evaluation of two datasets illustrates the potential for effective inconsistency detection and fixing of large KGs.Keywords
This publication has 46 references indexed in Scilit:
- Assessing the Evolution of LLM Capabilities for Knowledge Graph Engineering in 20232025
- Repairing SHACL Constraint Violations Using Answer Set Programming2022
- Implementing Large-Scale ABox Materialization Using Subgraph Reasoning2022
- Optimal ABox Repair w.r.t. Static $\mathcal {EL}$ TBoxes: From Quantified ABoxes Back to ABoxes2022
- Analysing Large Inconsistent Knowledge Graphs Using Anti-patterns2021
- Making Decisions with Knowledge Base Repairs2019
- OntoDebug: Interactive Ontology Debugging Plug-in for Protégé2018
- Fast ABox Consistency Checking Using Incomplete Reasoning and Caching2017
- Fast Approximate A-Box Consistency Checking Using Machine Learning2016
- Serving DBpedia with DOLCE – More than Just Adding a Cherry on Top2015