HiSem-RAG: A Hierarchical Semantic-Driven Retrieval-Augmented Generation Method
Source abstract
Abstract
Traditional retrieval-augmented generation (RAG) methods struggle with hierarchical documents, often causing semantic fragmentation, structural loss, and inefficient retrieval due to fixed strategies. To address these challenges, this paper proposes HiSem-RAG, a hierarchical semantic-driven RAG method. It comprises three key modules: (1) hierarchical semantic indexing, which preserves boundaries and relationships between sections and paragraphs to reconstruct document context; (2) a bidirectional semantic enhancement mechanism that incorporates titles and summaries to facilitate two-way information flow; and (3) a distribution-aware adaptive threshold strategy that dynamically adjusts retrieval scope based on similarity distributions, balancing accuracy with computational efficiency. On the domain-specific EleQA dataset, HiSem-RAG achieves 82.00% accuracy, outperforming HyDE and RAPTOR by 5.04% and 3.98%, respectively, with reduced computational costs. On the LongQA dataset, it attains a ROUGE-L score of 0.599 and a BERT_F1 score of 0.839. Ablation studies confirm the complementarity of these modules, particularly in long-document scenarios.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Information Retrieval and Search Behavior · Topic Modeling · Handwritten Text Recognition Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Large-Scale Stream Data Integration and Analysis Technology Beijing Key Laboratory, School of Artificial Intelligence and Computer Science, North China University of Technology, No. 5 Jinyuanzhuang Road, Beijing 100144, China
Funding
National Natural Science Foundation of China
References (45)
Show source-provided references
1. Retrieval-augmented generation for knowledge-intensive nlp tasks
2. Karpukhin, V., Oguz, B., Min, S., Lewis, P.S., Wu, L., Edunov, S., Chen, D., and Yih, W.-t. (2020, January 16–20). Dense Passage Retrieval for Open-Domain Question Answering. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), Online.10.18653/v1/2020.emnlp-main.550 ↗
3. Sawarkar, K., Mangal, A., and Solanki, S.R. (2024, January 7–9). Blended rag: Improving rag (retriever-augmented generation) accuracy with semantic search and hybrid query-based retrievers. Proceedings of the 2024 IEEE 7th International Conference on Multimedia Information Processing and Retrieval (MIPR), San Jose, CA, USA.10.1109/MIPR62202.2024.00031 ↗
4. Asai, A., Wu, Z., Wang, Y., Sil, A., and Hajishirzi, H. (2024, January 7–11). Self-rag: Learning to retrieve, generate, and critique through self-reflection. Proceedings of the Twelfth International Conference on Learning Representations, Vienna, Austria.
5. Mansurova, A., Mansurova, A., and Nugumanova, A. (2024). QA-RAG: Exploring LLM reliance on external knowledge. Big Data Cogn. Comput., 8.10.3390/bdcc8090115 ↗
6. Sarthi, P., Abdullah, S., Tuli, A., Khanna, S., Goldie, A., and Manning, C.D. (2024, January 7–11). Raptor: Recursive abstractive processing for tree-organized retrieval. Proceedings of the Twelfth International Conference on Learning Representations, Vienna, Austria.
7. Retrieval augmented generation-driven information retrieval and question answering in construction management10.1016/j.aei.2025.103158 ↗
8. Edge, D., Trinh, H., Cheng, N., Bradley, J., Chao, A., Mody, A., Truitt, S., Metropolitansky, D., Ness, R.O., and Larson, J. (2024). From local to global: A graph rag approach to query-focused summarization. arXiv.
9. Gao, F., Xu, S., Hao, W., and Lu, T. (2025). KA-RAG: Integrating Knowledge Graphs and Agentic Retrieval-Augmented Generation for an Intelligent Educational Question-Answering Model. Appl. Sci., 15.10.3390/app152312547 ↗
10. GenKP: Generative knowledge prompts for enhancing large language models10.1007/s10489-025-06318-3 ↗
11. Injury degree appraisal of large language model based on retrieval-augmented generation and deep learning10.1016/j.ijlp.2025.102070 ↗
12. Knowledge graph enhanced retrieval-augmented generation for failure mode and effects analysis
13. ALBERT with Knowledge Graph Encoder Utilizing Semantic Similarity for Commonsense Question Answering10.32604/iasc.2023.032783 ↗
14. Evaluating the ideal chunk size for a rag system using llamaindex
15. Xu, P., Ping, W., Wu, X., McAfee, L., Zhu, C., Liu, Z., Subramanian, S., Bakhturina, E., Shoeybi, M., and Catanzaro, B. (2024, January 7–11). Retrieval meets long context large language models. Proceedings of the Twelfth International Conference on Learning Representations, Vienna, Austria.
16. Yang, S. (2023, November 04). Advanced Rag 01: Small-to-Big Retrieval. Available online: https://medium.com/data-science/advanced-rag-01-small-to-big-retrieval-172181b396d4.
17. Krassovitskiy, A., Mussabayev, R., and Yakunin, K. (2025). LLM-Enhanced Semantic Text Segmentation. Appl. Sci., 15.10.3390/app151910849 ↗
18. Zhao, J., Ji, Z., Feng, Y., Qi, P., Niu, S., Tang, B., Xiong, F., and Li, Z. (2024). Meta-chunking: Learning efficient text segmentation via logical perception. arXiv.
19. Wang, K., Reimers, N., and Gurevych, I. (2023). DAPR: A benchmark on document-aware passage retrieval. arXiv.10.18653/v1/2024.acl-long.236 ↗
20. ChatGPT chemistry assistant for text mining and the prediction of MOF synthesis10.1021/jacs.3c05819 ↗
21. A Retrieval-Augmented Generation System for Large Language Models Based on Sliding Window Strategy
22. Gao, L., Ma, X., Lin, J., and Callan, J. (2023, January 9–14). Precise zero-shot dense retrieval without relevance labels. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), Toronto, ON, Canada.10.18653/v1/2023.acl-long.99 ↗
23. Peng, W., Li, G., Jiang, Y., Wang, Z., Ou, D., Zeng, X., Xu, D., Xu, T., and Chen, E. (2024, January 13–17). Large language model based long-tail query rewriting in taobao search. Proceedings of the Companion Proceedings of the ACM Web Conference 2024, Singapore.10.1145/3589335.3648298 ↗
24. Demystifying the black box: AI-enhanced logistic regression for lead scoring10.1007/s10489-025-06430-4 ↗
25. Tupayachi, J., and Li, X. (November, January 29). Conversational Geographic Question Answering for Route Optimization: An LLM and Continuous Retrieval-Augmented Generation Approach. Proceedings of the 17th ACM SIGSPATIAL International Workshop on Computational Transportation Science GenAI and Smart Mobility Session, Atlanta, GA, USA.
26. A Multi-Strategy Retrieval-Augmented Generation Method for Knowledge-Based Question Answering in the Military Domain
27. Ma, X., Gong, Y., He, P., Zhao, H., and Duan, N. (2023, January 6–10). Query rewriting in retrieval-augmented large language models. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Singapore.10.18653/v1/2023.emnlp-main.322 ↗
28. Kim, G., Kim, S., Jeon, B., Park, J., and Kang, J. (2023, January 6–10). Tree of clarifications: Answering ambiguous questions with retrieval-augmented large language models. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, Singapore.10.18653/v1/2023.emnlp-main.63 ↗
29. Yao, S., Zhao, J., Yu, D., Du, N., Shafran, I., Narasimhan, K., and Cao, Y. (2023). React: Synergizing reasoning and acting in language models. arXiv.
30. Construction and Evaluation of an Intelligent Question Answering System for Electric Power Knowledge Base Based on Large Language Models
31. Application of retrieval-augmented generation for interactive industrial knowledge management via a large language model10.1016/j.csi.2025.103995 ↗
32. Improving knowledge management in building engineering with hybrid retrieval-augmented generation framework10.1016/j.jobe.2025.112189 ↗
33. Empowering LLMs by hybrid retrieval-augmented generation for domain-centric Q&A in smart manufacturing10.1016/j.aei.2025.103212 ↗
34. DF-RAG: A Retrieval-Augmented Generation Method Based on Query Rewriting and Knowledge Selection
35. Multi-objective math problem generation using large language model through an adaptive multi-level retrieval augmentation framework10.1016/j.inffus.2025.103037 ↗
36. RAICL-DSC: Retrieval-Augmented In-Context Learning for Dialogue State Correction10.1016/j.knosys.2025.113423 ↗
37. Improved Retrieval Augmentation and LLM Chain-of-Thought for Maintenance Strategy Generation
38. Glass, M., Rossiello, G., Chowdhury, M.F.M., Naik, A.R., Cai, P., and Gliozzo, A. (2022). Re2G: Retrieve, rerank, generate. arXiv.10.18653/v1/2022.naacl-main.194 ↗
39. Rankrag: Unifying context ranking with retrieval-augmented generation in llms
40. Large language model for interpreting research policy using adaptive two-stage retrieval augmented fine-tuning method10.1016/j.eswa.2025.127330 ↗
41. Xu, C., Zhao, D., Wang, B., and Xing, H. (2024, January 5–8). Enhancing Retrieval-Augmented LMs with a Two-Stage Consistency Learning Compressor. Proceedings of the International Conference on Intelligent Computing, Tianjin, China.10.1007/978-981-97-5669-8_42 ↗
42. Wang, H., Wei, J., Jing, H., Song, H., and Xu, B. (2024). Meta-RAG: A Metadata-Driven Retrieval-Augmented Generation Framework for the Electric Power Domain. Comput. Eng., 1–11.
43. GLM, T., Zeng, A., Xu, B., Wang, B., Zhang, C., Yin, D., Zhang, D., Rojas, D., Feng, G., and Zhao, H. (2024). Chatglm: A family of large language models from glm-130b to glm-4 all tools. arXiv.
44. Chen, J., Xiao, S., Liu, P., Zhang, K., Lian, D., Xie, X., and Li, D. (2024). M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation. arXiv.10.18653/v1/2024.findings-acl.137 ↗
45. Zhang, T., Kishore, V., Wu, F., Weinberger, K.Q., and Artzi, Y. (2019). Bertscore: Evaluating text generation with bert. arXiv.