Elamur Research⌕
☰
home / papers / 2026 / FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow

FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow

Paper · 2026 · Bihao Zhan, Zongsheng Cao, Jie Zhou et al.

Summary

Graph-based RAG systems usually start retrieval from entities and spread relevance implicitly through the graph. This fails in two ways. Abstract queries with few explicit entity mentions may not activate the right starting nodes. Noisy activations can also push multi-hop traversal toward unrelated chains. FlowRAG addresses both problems.

Offline, it builds a heterogeneous graph with four node types: passages, LLM-generated passage summaries, sentences and named entities. The summaries act as coarse thematic hubs. At query time, entities are activated through two channels: matching the query to summaries and matching it to sentences. Relevance then flows along entity–passage edges, each weighted by how often the entity occurs in that passage relative to all entity mentions there. Weak links are pruned, and the strongest paths are handed to the generator as an explicit reasoning outline.

On HotpotQA, 2WikiMultiHopQA, MuSiQue and a medical benchmark, the authors report state-of-the-art results over strong GraphRAG baselines such as LinearRAG. The method matters because it makes multi-hop retrieval more interpretable and more robust to noise.

Key points

  • Graph structure: The graph has passage, summary, sentence and entity layers. The passage–summary and passage–entity matrices carry weights, while the summary–entity and sentence–entity matrices stay binary. Sparse storage keeps graph construction linearly scalable.
  • Edge weights: Passage–entity weights are within-passage normalized term frequency, which replaces LinearRAG's binary containment links. Entities come from lightweight NER (spaCy).
  • Model scaling: Results improve further when a stronger generator (GPT-4o) is used.
  • Case study: On a 2WikiMultihopQA kinship question, LinearRAG drifts to the famous Archduke Franz Ferdinand. FlowRAG instead follows a path through the intermediate entity Sophie and correctly returns Maximilian, Duke of Hohenberg.
  • Limitations: LLM summarization and graph propagation add extra compute. Frequency-based pruning may underweight rare but crucial domain terms. The framework currently supports only static knowledge bases.

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

🔬Shanghai Artificial Intelligence LaboratoryShanghai · China152 works · 13,166 citations🎓East China Normal UniversityShanghai · China69 works · 2,490 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 worksInformation Systems · 1,730 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,685 works

Cite

Cite as Bihao Zhan, Zongsheng Cao, Jie Zhou, Bo Zhang and Liang He. 2026. FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow. 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.1050

@inproceedings{zhan2026flowrag,
  title = {{FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow}},
  author = {Bihao Zhan, Zongsheng Cao, Jie Zhou, Bo Zhang and Liang He},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1050},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1050}
}

References

26 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2024. Self-rag: Learning to retrieve, generate, and critique through self-reflection. ICLR.
  2. Boyu Chen, Zirui Guo, Zidan Yang, Yuluo Chen, Junze Chen, Zhenghao Liu, Chuan Shi, and Cheng Yang. 2025. Pathrag: Pruning graph-based retrieval augmented generation with relational paths. Preprint, arXiv:2502.14902. — [arXiv:2502.14902](https://arxiv.org/abs/2502.14902) find
  3. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. 2025. From local to global: A graph rag approach to query-focused summarization. Preprint, arXiv:2404.16130. find
  4. Zirui Guo, Lianghao Xia, Yanhua Yu, Tu Ao, and Chao Huang. 2024. Lightrag: Simple and fast retrieval-augmented generation. arXiv preprint arXiv:2410.05779.
  5. Bernal Jiménez Gutiérrez, Yiheng Shu, Weijian Qi, Sizhe Zhou, and Yu Su. 2025. From rag to memory: Non-parametric continual learning for large language models. Preprint, arXiv:2502.14802.
  6. Haoyu Han, Li Ma, Harry Shomer, Yu Wang, Yongjia Lei, Kai Guo, Zhigang Hua, Bo Long, Hui Liu, Charu C Aggarwal, and 1 others. 2025. Rag vs. graphrag: A systematic evaluation and key insights. arXiv preprint arXiv:2502.11371. — [arXiv:2502.11371](https://arxiv.org/abs/2502.11371) find
  7. Xiaoxin He, Yijun Tian, Yifei Sun, Nitesh V. Chawla, Thomas Laurent, Yann LeCun, Xavier Bresson, and Bryan Hooi. 2024. G-retriever: Retrieval-augmented generation for textual graph understanding and question answering. Preprint, arXiv:2402.07630.
  8. Xanh Ho, Anh-Khoa Duong Nguyen, Saku Sugawara, and Akiko Aizawa. 2020. Constructing a multi-hop qa dataset for comprehensive evaluation of reasoning steps. arXiv preprint arXiv:2011.01060.
  9. Bernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. 2024. Hipporag: Neurobiologically inspired long-term memory for large language models. Advances in Neural Information Processing Systems, 37:59532–59569.
  10. Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick SH Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. Dense passage retrieval for open-domain question answering. In EMNLP (1), pages 6769–6781.
  11. 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–…
  12. Linhao Luo, Zicheng Zhao, Gholamreza Haffari, Dinh Phung, Chen Gong, and Shirui Pan. 2025. Gfm-rag: graph foundation model for retrieval augmented generation. arXiv preprint arXiv:2502.01113. — [arXiv:2502.01113](https://arxiv.org/abs/2502.01113) find
  13. Jose L Martinez-Rodriguez, Ivan López-Arévalo, and Ana B Rios-Alvarado. 2018. Openie-based approach for knowledge graph construction from text. Expert Systems with Applications, 113:339–355. find
  14. 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.
  15. Zhihong Shao, Yeyun Gong, Yelong Shen, Minlie Huang, Nan Duan, and Weizhu Chen. 2023. Enhancing retrieval-augmented large language models with iterative retrieval-generation synergy. arXiv preprint arXiv:2305.15294.
  16. Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, and Tie-Yan Liu. 2020. Mpnet: Masked and permuted pre-training for language understanding. Advances in neural information processing systems, 33:16857– 16867.
  17. Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. 2022. musique: Multi-hop questions via single-hop question composition. Transactions of the Association for Computational Linguistics, 10:539–554. find
  18. Liang Wang, Nan Yang, Xiaolong Huang, Binxing Jiao, Linjun Yang, Daxin Jiang, Rangan Majumder, and Furu Wei. 2022. Text embeddings by weakly-supervised contrastive pre-training. arXiv preprint arXiv:2212.03533.
  19. Wenhui Wang, Furu Wei, Li Dong, Hangbo Bao, Nan Yang, and Ming Zhou. 2020. Minilm: Deep self-attention distillation for task-agnostic compression of pre-trained transformers. Advances in neural information processing systems, 33:5776–5788. find
  20. Yu Wang, Nedim Lipka, Ryan A Rossi, Alexa Siu, Ruiyi Zhang, and Tyler Derr. 2024. Knowledge graph prompting for multi-document question answering. In Proceedings of the AAAI conference on artificial intelligence, volume 38, pages 19206–19214. find
  21. Zhishang Xiang, Chuanjie Wu, Qinggang Zhang, Shengyuan Chen, Zijin Hong, Xiao Huang, and Jinsong Su. 2025. When to use graphs in rag: A comprehensive analysis for graph retrieval-augmented generation. arXiv preprint arXiv:2506.05690. — [arXiv:2506.05690](https://arxiv.org/abs/2506.05690) find
  22. Shitao Xiao, Zheng Liu, Peitian Zhang, Niklas Muennighoff, Defu Lian, and Jian-Yun Nie. 2024. C-pack: Packed resources for general chinese embeddings. In Proceedings of the 47th international ACM SIGIR conference on research and development in information retrieval, pages 641–649.
  23. Shi-Qi Yan, Jia-Chen Gu, Yun Zhu, and Zhen-Hua Ling. 2024. Corrective retrieval augmented generation. Preprint, arXiv:2401.15884.
  24. Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William Cohen, Ruslan Salakhutdinov, and Christopher D Manning. 2018. Hotpotqa: A dataset for diverse, explainable multi-hop question answering. In Proceedings of the 2018 conference on empirical methods in natural language processing, pages 2369–…
  25. Yibo Zhao, Jiapeng Zhu, Ye Guo, Kangkang He, and Xiang Li. 2025. E 2 graphrag : Streamlininggraph − basedrag f orhighe f f iciencyande f f ectiveness. Preprint, arXiv:2505.24226. — [arXiv:2505.24226](https://arxiv.org/abs/2505.24226) find
  26. Luyao Zhuang, Shengyuan Chen, Yilin Xiao, Huachi Zhou, Yujing Zhang, Hao Chen, Qinggang Zhang, and Xiao Huang. 2025. Linearrag: Linear graph retrieval augmented generation on large-scale corpora. Preprint, arXiv:2510.10114. — [arXiv:2510.10114](https://arxiv.org/abs/2510.10114) find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhan-2026-flowrag-synergizing-explicit-reasoning-frequency/