Elamur Research⌕
☰
home / papers / 2026 / TagRAG: Tag-guided Hierarchical Knowledge Graph Retrieval-Augmented Generation

TagRAG: Tag-guided Hierarchical Knowledge Graph Retrieval-Augmented Generation

Paper · 2026 · Wenbiao Tao, Xinyuan Li, Yunshi Lan et al.

Summary

TagRAG addresses weaknesses of graph-based retrieval-augmented generation on query-focused summarization, where answers require a global view of a whole corpus. GraphRAG supports such reasoning, but it is costly. It needs many calls to large LLMs, performs poorly with smaller models, and must rebuild the full graph when new documents arrive. TagRAG works in two stages.

  1. Construction. An LLM extracts keyword-like object tags and their relations from document chunks. These tags are attached to a predefined root domain tag, forming hierarchical domain tag chains linked by "has subdomain" edges. Knowledge from the object tags is then summarized at the domain-tag level.
  2. Inference. The system retrieves domain-centric tags, gathers their chains and fused summaries, and generates a global answer.

The authors evaluated TagRAG on UltraDomain (Agriculture, CS, Legal, Mix), mainly with Qwen3-4B. It beat the baselines in most pairwise LLM-judged comparisons and was much cheaper to build and query than GraphRAG. Because new knowledge is merged into existing tag chains rather than rebuilt, the approach suits small-model deployments and corpora that grow over time.

Key points

  • Across UltraDomain Agriculture, CS, Legal and Mix, TagRAG has a 78.36% average winning rate over baselines. Compared with GraphRAG, it offers about 14.6× faster graph construction and 1.9× faster retrieval.
  • The baselines include Naive RAG, GraphRAG, LightRAG and MiniRAG. Answers are judged on Comprehensiveness, Diversity, Empowerment and Overall quality.
  • Incremental test: a graph was built step by step from 10 UltraDomain CS documents. TagRAG kept average winning rates above 80% against Naive RAG, GraphRAG and LightRAG. MiniRAG tied with it in the first round but fell behind as more documents were added.
  • Cross-domain update: a CS document was added to the Mix graph. TagRAG reached an Overall win rate of 58.0% (after vs. before the update), compared with 44.0% for GraphRAG. Its incremental construction took 6.37 h, versus 30.47 h for GraphRAG.
  • Robustness and limitations: the advantage held with bge-large-, bge-base- and bge-small-en-v1.5 retrievers. Indexing still depends on LLM calls, and the method does not handle images or video.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓East China Normal UniversityShanghai · China50 works · 2,174 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksComputer Vision and Pattern Recognition · 2,869 worksInformation Systems · 1,437 works

🧭 Topics 2Research subjects and languages the work is about.Information retrieval & search · 473 worksKnowledge graphs & ontologies · 192 works

Cite

Cite as Wenbiao Tao, Xinyuan Li, Yunshi Lan and Weining Qian. 2026. TagRAG: Tag-guided Hierarchical Knowledge Graph Retrieval-Augmented Generation. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.321

@inproceedings{tao2026tagrag,
  title = {{TagRAG: Tag-guided Hierarchical Knowledge Graph Retrieval-Augmented Generation}},
  author = {Wenbiao Tao, Xinyuan Li, Yunshi Lan and Weining Qian},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.321},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.321}
}

References

26 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Mariam Barry, Gaëtan Caillaut, Pierre Halftermeyer, Raheel Qader, Mehdi Mouayad, Dimitri Cariolaro, Fabrice Le Deit, and Joseph Gesnouin. 2025. Graphrag: leveraging graph-based efficiency to minimize hallucinations in llm-driven rag for finance data. In 31st International conference on Computational… find
  2. Boyu Chen, Zirui Guo, Zidan Yang, Yuluo Chen, Junze Chen, Zhenghao Liu, Chuan Shi, and Cheng Yang. 2025a. Pathrag: Pruning graph-based retrieval augmented generation with relational paths. arXiv preprint arXiv:2502.14902. — [arXiv:2502.14902](https://arxiv.org/abs/2502.14902) find
  3. Qinwen Chen, Wenbiao Tao, Zhiwei Zhu, Mingfan Xi, Liangzhong Guo, Yuan Wang, Wei Wang, and Yunshi Lan. 2025b. Comrag: Retrieval-augmented generation with dynamic vector stores for real-time community question answering in industry. In Proceedings of the 63rd Annual Meeting of the Association for Com… find
  4. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. 2024. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130. find
  5. Tianyu Fan, Jingyuan Wang, Xubin Ren, and Chao Huang. 2025. Minirag: Towards extremely simple retrieval-augmented generation. arXiv preprint arXiv:2501.06713. — [arXiv:2501.06713](https://arxiv.org/abs/2501.06713) find
  6. Wenqi Fan, Yujuan Ding, Liangbo Ning, Shijie Wang, Hengyun Li, Dawei Yin, Tat-Seng Chua, and Qing Li. 2024. A survey on rag meeting llms: Towards retrieval-augmented large language models. In Proceedings of the 30th ACM SIGKDD conference on knowledge discovery and data mining, pages 6491– 6501.
  7. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, Amy Yang, Angela Fan, Anirudh Goyal, Anthony Hartshorn, Aobo Yang, Archi Mitra, Archie Sravankumar, Artem Korenev, Arthur Hinsvark, and 542 o… — The Llama 3 Herd of Models
  8. Zirui Guo, Lianghao Xia, Yanhua Yu, Tu Ao, and Chao Huang. 2025. LightRAG: Simple and fast retrieval-augmented generation. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 10746–10761.
  9. Yubin Hong, Chaofan Li, Jingyi Zhang, and Yingxia Shao. 2025. Fg-rag: Enhancing query-focused summarization with context-aware fine-grained graph rag. arXiv preprint arXiv:2504.07103. — [arXiv:2504.07103](https://arxiv.org/abs/2504.07103) find
  10. Haoyu Huang, Yongfeng Huang, Junjie Yang, Zhenyu Pan, Yongqiang Chen, Kaili Ma, Hongzhi Chen, and James Cheng. 2025a. Retrieval-augmented generation with hierarchical knowledge. arXiv preprint arXiv:2503.10150. find
  11. Yiqian Huang, Shiqi Zhang, and Xiaokui Xiao. 2025b. Ket-rag: A cost-efficient multi-granular indexing framework for graph-rag. arXiv preprint arXiv:2502.09304. — [arXiv:2502.09304](https://arxiv.org/abs/2502.09304) find
  12. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, and 1 others. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276.
  13. Yunshi Lan, Xinyuan Li, Hanyue Du, Xuesong Lu, Ming Gao, Weining Qian, and Aoying Zhou. 2025. Survey of natural language processing for education: Taxonomy, systematic review, and future trends. IEEE Transactions on Knowledge and Data Engineering. find
  14. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, and 1 others. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in neural information processing systems, 33:9459–…
  15. Yuanyuan Liang, Keren Tan, Tingyu Xie, Wenbiao Tao, Siyuan Wang, Yunshi Lan, and Weining Qian. 2024. Aligning large language models to a domain-specific graph database for nl2gql. In Proceedings of the 33rd ACM international conference on information and knowledge management, pages 1367–1377. find
  16. Boci Peng, Yun Zhu, Yongchao Liu, Xiaohe Bo, Haizhou Shi, Chuntao Hong, Yan Zhang, and Siliang Tang. 2024. Graph retrieval-augmented generation: A survey. arXiv preprint arXiv:2408.08921.
  17. Hongjin Qian, Zheng Liu, Peitian Zhang, Kelong Mao, Defu Lian, Zhicheng Dou, and Tiejun Huang. 2025. Memorag: Boosting long context processing with global memory-enhanced retrieval augmentation. In Proceedings of the ACM on Web Conference 2025, pages 2366–2377. find
  18. Parth Sarthi, Salman Abdullah, Aditi Tuli, Shubh Khanna, Anna Goldie, and Christopher D Manning. 2024. Raptor: Recursive abstractive processing for tree-organized retrieval. In The Twelfth International Conference on Learning Representations.
  19. Wenbiao Tao, Hanlun Zhu, Keren Tan, Jiani Wang, Yuanyuan Liang, Huihui Jiang, Pengcheng Yuan, and Yunshi Lan. 2024. Finqa: A training-free dynamic knowledge graph question answering system in finance with llm-based revision. In Joint European Conference on Machine Learning and Knowledge Discovery in… find
  20. Qwen Team. 2025. Qwen3 technical report. Preprint, arXiv:2505.09388.
  21. Shu Wang, Yixiang Fang, Yingli Zhou, Xilin Liu, and Yuchi Ma. 2025. Archrag: Attributed community-based hierarchical retrieval-augmented generation. arXiv preprint arXiv:2502.09891. — [arXiv:2502.09891](https://arxiv.org/abs/2502.09891) find
  22. Nirmalie Wiratunga, Ramitha Abeyratne, Lasal Jayawardena, Kyle Martin, Stewart Massie, Ikechukwu Nkisi-Orji, Ruvan Weerasinghe, Anne Liret, and Bruno Fleisch. 2024. Cbr-rag: case-based reasoning for retrieval augmented generation in llms for legal question answering. In International Conference on C…
  23. Shitao Xiao, Zheng Liu, Peitian Zhang, and Niklas Muennighoff. 2023. C-pack: Packaged resources to advance general chinese embedding. Preprint, arXiv:2309.07597.
  24. Yaoze Zhang, Rong Wu, Pinlong Cai, Xiaoman Wang, Guohang Yan, Song Mao, Ding Wang, and Botian Shi. 2026. Leanrag: Knowledge-graph-based generation with semantic aggregation and hierarchical retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 40, pages 34862– 34869. find
  25. Xuejiao Zhao, Siyan Liu, Su-Yin Yang, and Chunyan Miao. 2025. Medrag: Enhancing retrieval-augmented generation with knowledge graph-elicited reasoning for healthcare copilot. In Proceedings of the ACM on Web Conference 2025, pages 4442–4457. find
  26. Zulun Zhu, Tiancheng Huang, Kai Wang, Junda Ye, Xinghe Chen, and Siqiang Luo. 2025. Graph-based approaches and functionalities in retrieval-augmented generation: A comprehensive survey. ACM Computing Surveys. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/tao-2026-tagrag-tag-guided-hierarchical-knowledge/