A HIERARCHICAL SEMANTIC KNOWLEDGE GRAPH MODEL FOR MULTILEVEL KNOWLEDGE REPRESENTATION
DOI:
https://doi.org/10.35546/kntu2078-4481.2025.4.3.10Keywords:
knowledge graph, semantic knowledge graph, semantic metrics, embedding, clustering.Abstract
Knowledge graphs are a fundamental technology for structured knowledge representation, logical inference and semantic operations in complex information systems. However, widely used semantic knowledge graph models and knowledge graphs augmented with large language models operate at a fixed level of semantic detail and do not support multi-level abstraction. Hierarchical, multiscale knowledge graph models allow constructing topological hierarchies based on meso- and macro-level structures, but their semantic properties and possible data operations require additional research. This paper introduces a hierarchical semantic knowledge graph (HSKG) model defined as a sequence of levels with their own sets of entities and relations, enabling multi-level knowledge representation. We formally specify the structure of HSKG, a hybrid semantic connectivity metric that combines cosine similarity of multi-dimensional embeddings with the Jaccard measure on graph neighborhoods, and an iterative hierarchical aggregation algorithm that groups entities into level-clusters. Based on these components, we implement a software prototype and conduct an experimental evaluation on subsets of the DbPedia50 and YAGO3-10 benchmark datasets with pre-generated multi-dimensional embeddings. The results demonstrate that HSKG reduces the average response time of semantic queries (by up to 30–80 % compared to single-level models and Louvain/Leiden-based clustering models) and yields clusters with optimized semantic connectivity and internal distribution compactness. These findings confirm the suitability of the proposed model and related semantic operations for building scalable tools for semantic analysis, search and logical inference in dynamic knowledge bases, recommendation generation systems and intelligent assistants.
References
ESCARGOT: An AI agent leveraging large language models, dynamic graph of thoughts, and biomedical knowledge graphs for enhanced reasoning / N. Matsumoto et al. Bioinformatics, Volume 41, Issue 2, 2025. DOI: https://doi.org/10.1093/bioinformatics/btaf031.
Unifying Large Language Models and Knowledge Graphs: A Roadmap / S. Pan et al. IEEE Transactions on Knowledge and Data Engineering. 2024. P. 1–20. DOI: https://doi.org/10.1109/tkde.2024.3352100.
Large language models for intelligent RDF knowledge graph construction: results from medical ontology mapping / A. Mavridis et al. Frontier in Artificial Intelligence, Volume 8, No 1546179, 2025. DOI: https://doi.org/10.3389/frai.2025.1546179.
Zichao Li, Zong Ke, Puning Zhao. Injecting Structured Knowledge into LLMs via Graph Neural Networks. In Proceedings of the 1st Joint Workshop on Large Language Models and Structure Modeling (XLLM 2025), Vienna, Austria. Association for Computational Linguistics. С. 16–25. URL: https://aclanthology.org/2025.xllm-1.3.pdf
Large Language Models for Knowledge Graph Embedding: A Survey / B. Liu et al. Mathematics, Volume 13, Issue 14, 2025. DOI: https://doi.org/10.3390/math13142244.
Comprehensive survey on hierarchical clustering algorithms and the recent developments / X. Ran et al. Artificial Intelligence Review. 2022. DOI: https://doi.org/10.1007/s10462-022-10366-3.
GeOKG: geometry-aware knowledge graph embedding for Gene Ontology and genes / C. Jeong et al. Bioinformatics, Volume 41, Issue 4, 2025. DOI: https://doi.org/10.1093/bioinformatics/btaf160.
Detecting hallucinations in large language models using semantic entropy / S. Farquhar et al. Nature. 2024. Vol. 630, no. 8017. P. 625–630. DOI: https://doi.org/10.1038/s41586-024-07421-0.
Zihao Xu, Zhejun Shen, Qunzhi Zhou, Petar Ristoski. Knowledge Graph-Enhanced Retrieval Augmented Generation for E-Commerce / RAGE-KG 2024: Retrieval-Augmented Generation Enabled by Knowledge Graphs, November 11, 2024, Baltimore, Maryland. URL: https://2024.rage-kg.org/papers/RAGE-KG_2024_paper_6.pdf
Knowledge Graph Embedding for Link Prediction / A. Rossi et al. ACM Transactions on Knowledge Discovery from Data. 2021. Vol. 15, no. 2. P. 1–49. DOI: https://doi.org/10.1145/3424672.
Linders J., Tomczak J. M. Knowledge graph-extended retrieval augmented generation for question answering. Applied Intelligence. 2025. Vol. 55, no. 17. DOI: https://doi.org/10.1007/s10489-025-06885-5.
Zhong Z., Li C., Pang J. Hierarchical message-passing graph neural networks. Data Mining and Knowledge Discovery, Volume 37, 2023. P. 381-408. DOI: https://doi.org/10.1007/s10618-022-00890-9.
Smaili F. Z., Gao X., Hoehndorf R. Onto2Vec: joint vector-based representation of biological entities and their ontology-based annotations. Bioinformatics, Volume 34, Issue 13, 2018. P. i52–i60. DOI: https://doi.org/10.1093/bioinformatics/bty259.
Information retrieval using cosine and jaccard similarity measures in vector space model / A. Jain et al. International Journal of Computer Applications, Volume 164, No 6, 2017. P. 28-30. DOI: https://doi.org/10.5120/ijca2017913699.
Hierarchical clustering that takes advantage of both density-peak and density-connectivity / Y. Zhu et al. Information Systems. 2022. Vol. 103. P. 101871. URL: https://doi.org/10.1016/j.is.2021.101871.
ProjectHSKG: Project for Hierarchical Semantic Knowledge Graph core & experiments demonstration. URL: https://github.com/YANZAAN/ProjectHSKG (дата звернення: 18.11.2025).
Open-world knowledge graph completion for unseen entities and relations via attentive feature aggregation / B. Oh et al. Information Sciences. 2022. Vol. 586. P. 468–484. DOI: https://doi.org/10.1016/j.ins.2021.11.085.
Mahdisoltani F., Biega J., Suchanek F. M. YAGO3: A Knowledge Base from Multilingual Wikipedias. Proceedings of the 7th Conference on Innovative Data Systems Research, Asilomar, CA, USA, 2015. URL: https://imt.hal.science/hal-01699874v1.
Traag, V., Waltman, L., van Eck, N. From Louvain to Leiden: guaranteeing well-connected communities. Scientific Reports, Volume 9, Article number 5233, 2019. DOI: https://doi.org/10.1038/s41598-019-41695-z.
Information and analytical support of gender audit of educational management system / T. Fesenko et al. Bulletin of NTU "KhPI". Series: Strategic management, portfolio, program and project management. 2022. No. 1(5). P. 70–79. DOI: https://doi.org/10.20998/2413-3000.2022.5.9.
