Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68952
Volume:
10 |
11 |
Nov 2026
Pages:
1-11
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.

Computer Science
AKS University Satna
Computer Science
AKS University Satna
Computer Science
AKS University Satna
Retrieval-Augmented Generation (RAG) has emerged as an effective approach for enhancing the factual grounding of Large Language Models (LLMs) by incorporating external knowledge. However, conventional RAG systems often face challenges in capturing complex relationships, integrating distributed evidence, and supporting multi-hop reasoning, particularly for complex questions that need several connected facts. Knowledge Graph-Augmented Retrieval-Augmented Generation (KG-RAG) addresses these limitations by integrating organized knowledge representations with semantic retrieval and language generation. This review examines recent developments in KG-RAG approaches, focusing on architectures, retrieval strategies, knowledge integration mechanisms, reasoning techniques, and evaluation methodologies. Representative frameworks, including GraphRAG, HippoRAG, KG²RAG, LightRAG, and Think-on-Graph, are analyzed to identify their methodological contributions and limitations. The review presents a taxonomy of KG-RAG approaches based on knowledge graph construction, retrieval mechanisms, evidence integration, and answer generation. It also discusses benchmark datasets and evaluation metrics relevant to retrieval effectiveness, answer faithfulness, and multi-hop reasoning. The literature synthesis identifies major challenges involving knowledge graph incompleteness, evidence reliability, hybrid retrieval, computational scalability, and domain adaptation. Prospects for future research include confidence-aware evidence fusion, adaptive retrieval strategies, reasoning-path verification, and robust domain-specific evaluation. Overall, this review provides a structured understanding of KG-RAG developments and highlights opportunities for building more reliable, efficient, and context-aware question-answering systems.
KEYWORDS
This review examined recent dev elopments in Knowledge Graph-Augmented Retrieval-Augmented Generation (KG-RAG) for multi-hop question, focusing on existing architectures, retrieval strategies, knowledge integration techniques, evaluation methods, and research challenges. The reviewed studies demonstrate that integrating knowledge graphs with RAG provides a promising approach for supporting relationship-aware information retrieval, structured evidence organization, and multi-hop logic. Existing frameworks, including GraphRAG, HippoRAG, KG²RAG, LightRAG, and Think-on-Graph, demonstrate different approaches to combining structured knowledge with LLM. (Edge et al., 2024; Gutiérrez et al., 2024; Sun et al., 2024; Zhu et al., 2025; Guo et al., 2025). However, their effectiveness depends on factors such as knowledge graph quality, retrieval accuracy, reasoning complexity, and computational efficiency. The review identifies several important challenges, including incomplete knowledge graphs, unreliable multi-hop evidence, difficulties in integrating structured and unstructured information, scalability limitations, and domain adaptation. Addressing these challenges requires further investigation into adaptive hybrid retrieval, confidence-aware evidence fusion, reasoning-path verification, and efficient graph maintenance. Overall, KG-RAG represents a valuable direction for developing more dependable and context-aware question-answering systems. Future research should emphasize robust knowledge integration, standardized evaluation, and domain-specific adaptability to enhance the accuracy, efficiency, and trustworthiness of RAG.
[1] Zhu, X., Xie, Y., Liu, Y., Li, Y., & Hu, W. (2025, April). Knowledge graph-guided retrieval augmented generation. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers) (pp. 8912-8924). [2] Edge, D., Trinh, H., Cheng, N., Bradley, J., Chao, A., Mody, A., ... & Larson, J. (2024). From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130. [3] Gutiérrez, B. J., Shu, Y., Gu, Y., Yasunaga, M., & Su, Y. (2024). Hipporag: Neurobiologically inspired long-term memory for large language models. Advances in neural information processing systems, 37, 59532-59569. [4] Guo, Z., Xia, L., Yu, Y., Ao, T., & Huang, C. (2025, November). LightRAG: Simple and Fast Retrieval-Augmented Generation. In EMNLP (Findings) (pp. 10746-10761). [5] Shrivastava, R., & Gautam, C. S. (2026). Enhanced Heart Disease Prediction using Ensemble of Machine Learning Models. International Journal of Computer Applications, 187(104), 40-46. [6] He, X., Tian, Y., Sun, Y., Chawla, N. V., Laurent, T., LeCun, Y., ... & Hooi, B. (2024). G-retriever: Retrieval-augmented generation for textual graph understanding and question answering. Advances in Neural Information Processing Systems, 37, 132876-132907. [7] Linders, J., & Tomczak, J. M. (2025). Knowledge graph-extended retrieval augmented generation for question answering. Applied Intelligence, 55(17), 1102. [8] Zhou, D., Zhu, Y., Wang, X., He, Y., Chen, J., Staab, S., & Kharlamov, E. (2025). Evaluating knowledge graph based retrieval augmented generation methods under knowledge incompleteness. arXiv preprint arXiv:2504.05163. [9] Ma, S., Xu, C., Jiang, X., Li, M., Qu, H., Yang, C., ... & Guo, J. (2025, May). Think-on-graph 2.0: Deep and faithful large language model reasoning with knowledge-guided retrieval augmented generation. In International Conference on Learning Representations (Vol. 2025, pp. 52782-52806). [10] Sun, J., Xu, C., Tang, L., Wang, S., Lin, C., Gong, Y., ... & Guo, J. (2024, May). Think-on-graph: Deep and responsible reasoning of large language model on knowledge graph. In International Conference on Learning Representations (Vol. 2024, pp. 3868-3898).
Publication Date
11/13/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
5
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...