Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Youtu-GraphRAG: Vertically Unified Agents for Graph Retrieval-Augmented Complex Reasoning

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A vertically unified agentic GraphRAG design ties graph construction to retrieval through one evolving schema, reporting up to 90.71% lower token costs and 16.62% higher accuracy across six benchmarks.

desk verdict Serious GraphRAG systems paper with a useful new benchmark, but the core clustering math is under-specified and the headline numbers aren't directly traceable. read the letter →

arxiv 2508.19855 v3 pith:FCIHAAYF submitted 2025-08-27 cs.IR

Junnan Dong , Siyu An , Yifei Yu , Qian-Wen Zhang , Linhao Luo , Xiao Huang , Yunsheng Wu , Di Yin
show 1 more author
Xing Sun
This is my paper · ORCID
classification cs.IR
keywords GraphRAGschemacommunitydetectionmulti-hopquestionansweringagenticretrievalknowledgetreeleakageanonymousbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Youtu-GraphRAG sets out to prove that graph construction and graph retrieval in a GraphRAG system should be optimized as one connected process, not as separate components. It ties both stages to a single, evolving graph schema: the same schema bounds the extraction agent to relevant entity and relation types and later guides the retrieval agent in decomposing complex questions into parallel sub-queries. On six benchmarks the paper reports state-of-the-art top-20 accuracy while cutting graph-construction token costs by up to 90.71% relative to strong baselines, and it adds two anonymized datasets with an Anonymity Reversion task so that accuracy gains measure retrieved evidence rather than the model's memorized knowledge. If the claim holds, GraphRAG systems can become cheaper and more accurate at the same time, and moving to a new domain reduces to seeding a small schema rather than re-engineering the pipeline.

What carries the argument

The load-bearing object is the graph schema S=(Se,Sr,Sattr), the set of targeted entity types, relation types, and attribute types that acts as the single interface between construction and retrieval. It bounds extraction to schema-compliant triples, is expanded adaptively via a confidence threshold, and is then read again by the retriever to decompose queries into sub-queries that match valid graph patterns. The second mechanism is dually-perceived community detection, a three-stage procedure that encodes each entity by aggregating embeddings of its one-hop triples, initializes clusters with K-means, and then iteratively fuses communities using a dual-perception score blending relational Ja

What would settle it

Re-run the full pipeline on all six benchmarks with one fixed value of the community-merging threshold and no per-dataset tuning. If top-20 accuracy under reject mode collapses on any dataset, the hierarchical knowledge tree is not domain-robust as claimed; this directly tests whether the merging rule defines a meaningful, transferable criterion.

Watch

Extended reading notes

Core claim

The central claim is that a vertically unified agentic paradigm—rather than isolated improvements to graph construction or retrieval—is what drives both cost savings and reasoning gains in GraphRAG. The discovery is the shared graph schema itself, deployed twice. During construction it constrains the extraction agent to targeted entity types, relations, and attribute types and expands automatically with high-confidence patterns; during retrieval the same schema is interpreted by an agentic retriever to split a complex query into schema-aligned atomic sub-queries and to pick among entity, triple, community, and path-traversal retrieval routes. A second mechanism, dually-perceived community de

Load-bearing premise

The central claim depends on a community-merging rule in Eq. (7) that is never fully defined—the 'dual-perception divergence' it compares is never specified and its thresholds are never given values—so if that rule is ill-posed or tuned per dataset, the hierarchical knowledge tree that all downstream retrieval builds on has no guaranteed coherence.

Editorial extensions

If this is right

  • If the unified schema design is correct, graph construction and retrieval no longer trade off against each other: the same schema that slims extraction also tells the retriever where to look, so both accuracy and token cost can improve together.
  • The four-level knowledge tree gives one structure for both global and local questions, allowing top-down community filtering for broad queries and bottom-up triple and attribute retrieval for fact-level ones.
  • Schema expansion from document feedback means a new domain can be entered with a small seed schema and grown automatically, rather than requiring a hand-built ontology per corpus.
  • Because the anonymized evaluation forces reliance on retrieved context, the reported margins under reject mode indicate the retriever actually supplies complete evidence instead of letting the language model answer from memory.
  • An agentic retriever that decomposes queries along schema paths and reflects on its steps should scale to more compositional questions, since each sub-query is grounded in a valid graph pattern.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The schema could become a runtime control surface: editing entity or relation types at deployment time would change both extraction and retrieval behavior, giving practitioners a single knob for domain adaptation—an extension the paper does not test.
  • The Anonymity Reversion task could be applied to existing public benchmarks by re-anonymizing their entities, producing a low-cost way to re-baseline prior GraphRAG results and expose which systems relied on memorized knowledge.
  • The undefined merge criterion in Eq. (7) leaves room for a formal treatment: if the divergence is spelled out and the threshold derived rather than tuned, the community tree could be made provably coherent, turning the reported empirical gains into a transferable principle.
  • The schema-aligned decomposition suggests the same knowledge tree could support interactive, multi-turn query refinement without rebuilding the index, since re-planning sub-queries is cheaper than re-retrieving from scratch.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes Youtu-GraphRAG, a schema-guided GraphRAG system that unifies graph construction and retrieval. A seed graph schema bounds an LLM extraction agent; a 'dually-perceived' community detection method combines structural and semantic signals to build a four-level hierarchical knowledge tree; an agentic retriever decomposes complex queries into schema-aligned sub-queries with iterative reflection; and a new AnonyRAG dataset with an 'Anonymity Reversion' task is introduced to mitigate LLM knowledge leakage. The paper reports state-of-the-art answer accuracy on six benchmarks (HotpotQA, 2WikiMultiHopQA, MuSiQue, G-Bench, AnonyRAG-CHS, AnonyRAG-ENG) under two LLM backbones, together with large reductions in graph-construction token costs.

Significance. If the reported results are reproducible, the paper makes a useful contribution: the vertical integration of schema-guided extraction, hierarchical community construction, and schema-aware retrieval is a plausible route to improving both accuracy and token efficiency in GraphRAG. The proposed dual-mode evaluation (reject/open) and the anonymous dataset address a real weakness in existing GraphRAG benchmarks, namely knowledge leakage from pretrained LLMs. The authors also release code and data, which is a strength. However, the central construction algorithm is under-specified in the manuscript, and the experimental reporting does not currently support the strength of the headline claims. The core idea is defensible, but the paper needs substantial revision before the claims can be independently verified.

major comments (3)
  1. [§3.2, Eqs. (5)–(7)] The dual-perception community detection algorithm is not well-defined as written. Eq. (5) defines phi(e_i, C_m) = S_r ⊕_lambda S_s, and Eq. (6) then defines S_s(e_i, C_m) = phi(F_Theta(T_i), sum_j F_Theta(T_j)). Unless the right-hand phi in Eq. (6) is a placeholder for a different similarity (e.g., cosine), the definition is circular. The operator ⊕ and the weighting lambda are never specified, and no numeric value for lambda is reported. Eq. (7) merges clusters when E[phi(e_i,C_a)] - E[phi(e_i,C_b)] < epsilon, but the expectation is taken over no stated distribution, the expression is a signed difference rather than a divergence, and epsilon (like mu in Eq. (3)) is never given. Because every downstream retrieval route consumes the knowledge tree built on these communities, the core indexing step is not reproducible from the manuscript. The contribution text calls the method 'theoretical
  2. [§4.4, Tables 1–3] The empirical evaluation lacks basic statistical reporting. No error bars, confidence intervals, or number of independent runs are given for any table. The ablation study in Table 3 appears to be a single run, and the differences between the full model and ablations (e.g., 19.8 points on 2Wiki for 'w/o Agent') could be within run-to-run variance given the observed prompt sensitivity noted in §4.1. In addition, the LLM-as-judge for response similarity is DeepSeek-V3-0324, the same model family used for graph construction, retrieval generation, and the base LLM in half the experiments; this creates a risk of self-preference bias. Please report variance across multiple seeds/runs, justify the choice of judge, and define the 'top-20 accuracy' metric precisely (is it response-level exact/semantic-match accuracy, retrieval recall, or something else?).
  3. [Abstract and Figure 5/6] The headline claims of 'up to 90.71% saving of token costs and 16.62% higher accuracy' are not traceable to any table or figure. Figure 5 and Figure 6 are image-only with no numeric token-consumption table, and the 16.62% accuracy gain does not correspond directly to any pairwise comparison in Table 1 or Table 2. The reader cannot verify which dataset, mode, and baseline produced these numbers. Please add a table of construction-time/token costs for all methods and datasets, and state exactly how the headline percentages are computed.
minor comments (5)
  1. [Contributions and Abstract] The abstract says 'six challenging benchmarks' while the Contributions list says 'five challenging benchmarks.' The experiments use six; please make the count consistent.
  2. [Tables 1–2 and §4.2] The anonymized datasets are referred to as 'AnonyRAG-CHS' and 'AnonyRAG-ENG' in the text but appear as 'Annoy-CHS' and 'Annoy-ENG' in the tables. Use a single consistent name.
  3. [Eq. (6) and surrounding text] After Eq. (6), the text says 'Ss denotes the Jaccard similarity matrix computed over the multiset of incident relation types,' but Ss as defined in Eq. (6) is a semantic cosine similarity; the Jaccard term is Sr. The notation appears swapped.
  4. [Eq. (3)] The schema-update rule uses the notation fLLM(d, S) ⊙ S, but neither the output type of fLLM nor the ⊙ operation is defined. Please clarify.
  5. [Throughout] Several typos: 'Anonymous Revertion' in the Contributions section should be 'Anonymity Reversion'; 'HotptQA' in Table 3 should be 'HotpotQA'; 'vito role' in §4.6 should be 'vital role.'

Circularity Check

1 steps flagged · score 6.0 of 10

The dually-perceived community-detection score is self-referential as written: Eq. (5) defines φ via S_s, while Eq. (6) defines S_s via φ; otherwise the empirical evaluation is externally benchmarked and not circular.

  1. self definitional [Section 3.2, Eqs. (5)–(6)]
    "ϕ(ei, Cm) = Sr(ei, Cm) ⊕ λ Ss(ei, Cm), (5) with Sr(ei, Cm) = ∥Ψ(ei) ∩ Ψ(Cm)∥2 / ∥Ψ(ei) ∪ Ψ(Cm)∥2, Ss(ei, Cm) = ϕ(FΘ(Ti), ∑_{j∈Cm} FΘ(Tj)), (6) where Ss denotes the Jaccard similarity matrix computed over the multiset of incident relation types Ψ(·)."

    Equation (5) defines the dual-perception affinity φ(e_i,C_m) as a combination of relational overlap S_r and semantic similarity S_s. Equation (6) then defines S_s(e_i,C_m) by applying the same function φ to the entity embedding F_Θ(T_i) and the community aggregate Σ F_Θ(T_j). Thus φ is defined in terms of S_s and S_s in terms of φ, with no independent base case or fixed-point semantics supplied; the surrounding prose mentioning cosine similarity does not appear as a formal definition. This score drives centroid selection, the merging criterion in Eq. (7), and keyword selection in Eq. (8), so the construction of the four-level knowledge tree—the paper's central indexing contribution—is not well-defined by the displayed equations.

full rationale

The only genuine circular step found is the self-referential pair of Eqs. (5)–(6): the affinity score φ is defined using S_s, and S_s is defined using φ. As written, this makes the dually-perceived community detection—and therefore the entire knowledge-tree construction—non-computable from the manuscript. Underspecified elements such as the expectation in Eq. (7), the threshold ε, and the coefficient λ are reproducibility and correctness risks, but they are not themselves circular reductions. The paper's headline performance claims are empirical, benchmarked against external datasets (HotpotQA, MuSiQue, 2Wiki, G-Bench, AnonyRAG) with fixed base models, so they do not reduce to fitted constants or renamed predictions. Citations to the authors' own prior work (e.g., GraphRAG-Bench) are used as a dataset and background references, not as load-bearing uniqueness theorems or ansatz-justifying authorities; the 'theoretically-grounded' claim is asserted without proof, but absence of proof is not circularity. Score 6 reflects that the central indexing algorithm contains a definitional loop as written, even though the empirical comparison itself has independent content.

Assumptions & free parameters 6 free parameters · 5 assumptions · 2 invented entities

The central claims rest on hand-chosen hyperparameters (µ, λ, ε, β, η, d) and a hand-defined seed schema, plus several domain assumptions about LLM extraction, embedding quality, and the LLM-as-judge. The AnonyRAG dataset is an external artifact with independent evidence, but the schema itself is validated only within the paper.

free parameters (6)
  • µ (schema expansion confidence threshold)
    Controls acceptance of new schema elements in Eq. (3); value not disclosed, likely tuned.
  • λ (semantic weight in dual-perception score)
    Weights semantic vs relational similarity in Eq. (5); value not disclosed.
  • ε (community merge threshold)
    Determines when clusters merge in Eq. (7); value not disclosed and divergence undefined.
  • β (cluster granularity) = 10
    Sets approximate minimum of 10 entities per initial cluster; chosen by hand.
  • η (max initial clusters) = 200
    Prevents excessive fragmentation in K-means; chosen by hand.
  • d (DFS max depth) = 5
    Limits path traversal depth in retrieval; chosen by hand.
assumptions (5)
  • domain assumption A frozen LLM can reliably extract schema-compliant triples from documents.
    Invoked in Eq. (2) and §3.1; no human verification of extraction quality is reported.
  • domain assumption LLM embeddings of triples capture both structural and semantic similarity for clustering.
    Used in Eq. (4) and Eq. (5); the embedding quality is not benchmarked independently.
  • domain assumption The community merge criterion (Eq. 7) produces a meaningful hierarchy.
    The rule is under-specified; assumes differences in E[φ] indicate semantic divergence without formal definition.
  • domain assumption LLM-as-judge (DeepSeek-V3-0324) provides reliable accuracy scoring.
    Used in §4.1 for evaluating responses; no validation of the judge against human labels.
  • domain assumption Anonymization in AnonyRAG breaks pretrained memory shortcuts while preserving testable semantics.
    Appendix B shows zero-shot degradation, but no formal proof that all leakage is removed.
invented entities (2)
  • Seed graph schema (S)
    purpose: Bounds extraction and retrieval to targeted entity types, relation types, and attribute types.
    The schema is hand-initialized and auto-expanded; effectiveness is only shown via downstream benchmark accuracy, not independently validated.
  • AnonyRAG dataset independent evidence
    purpose: Evaluates GraphRAG under anonymized entities to mitigate knowledge leaking.
    The dataset is shipped and could be used by other researchers; zero-shot degradation on anonymized questions is a falsifiable handle.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Youtu-GraphRAG: Vertically Unified Agents for Graph Retrieval-Augmented Complex Reasoning." pith.science (2026). https://pith.science/paper/FCIHAAYF

@misc{pith2026250819855,
  author       = {Pith},
  title        = {Pith review of: Youtu-GraphRAG: Vertically Unified Agents for Graph Retrieval-Augmented Complex Reasoning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FCIHAAYF}},
  note         = {Machine review of arXiv:2508.19855}
}
read the original abstract

Graph retrieval-augmented generation (GraphRAG) has effectively enhanced large language models in complex reasoning by organizing fragmented knowledge into explicitly structured graphs. Prior efforts have been made to improve either graph construction or graph retrieval in isolation, yielding suboptimal performance, especially when domain shifts occur. In this paper, we propose a vertically unified agentic paradigm, Youtu-GraphRAG, to jointly connect the entire framework as an intricate integration. Specifically, (i) a seed graph schema is introduced to bound the automatic extraction agent with targeted entity types, relations and attribute types, also continuously expanded for scalability over unseen domains; (ii) To obtain higher-level knowledge upon the schema, we develop novel dually-perceived community detection, fusing structural topology with subgraph semantics for comprehensive knowledge organization. This naturally yields a hierarchical knowledge tree that supports both top-down filtering and bottom-up reasoning with community summaries; (iii) An agentic retriever is designed to interpret the same graph schema to transform complex queries into tractable and parallel sub-queries. It iteratively performs reflection for more advanced reasoning; (iv) To alleviate the knowledge leaking problem in pre-trained LLM, we propose a tailored anonymous dataset and a novel 'Anonymity Reversion' task that deeply measures the real performance of the GraphRAG frameworks. Extensive experiments across six challenging benchmarks demonstrate the robustness of Youtu-GraphRAG, remarkably moving the Pareto frontier with up to 90.71% saving of token costs and 16.62% higher accuracy over state-of-the-art baselines. The results indicate our adaptability, allowing seamless domain transfer with minimal intervention on schema.

Figures

Figures reproduced from arXiv: 2508.19855 by the authors.

Figure 1
Figure 1. A sketched comparison among existing pipelines and Youtu-GraphRAG. represents a non-tailored component, indicat￾ing current methods focus on either graph construction (a) or retrieval (b) in isolation, while Youtu-GraphRAG proposes a unified paradigm (c) for superior complex reasoning. memory and personalized PageRank algorithms for context￾aware retrieval. Second, in terms of graph construction, existing methods ca… view at source ↗
Figure 2
Figure 2. A toy overview of Youtu-GraphRAG that unifies graph construction and retrieval through a schema-guided agentic paradigm. (i) An extraction agent automatically processes documents into structured knowledge via targeted entity/relation extraction; (ii) A four-level knowledge tree is constructed upon the schema with a community detection that fuses topological structures and graph semantics, enabling hierarchical reaso… view at source ↗
Figure 3
Figure 3. An overview of our dually-perceived community detection. (a) Input graph partitioning into initial communi￾ties via triple embeddings; (b) community center identification through joint consideration of topology connectivity and subgraph semantic similarity; and (c) iterative pairwise community merging to form the final hierarchy. Distinct colors represent functionally coherent communities. and incorporate an adaptiv… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The figure contrasts three query-resolution strategies for a multi-hop question. While embedding matching retrieves disjointed facts (left) and traditional agents use repetitive templates (right), our agentic decomposer (center) leverages domain schema to plan efficien…
Figure 6
Figure 6. Figure 6: Youtu-GraphRAG effectively moves the Pareto fron￾tier with lower token costs and higher performance. While existing GraphRAG methods face a dilemma to balance the token consumption during con￾struction and the accuracy for final generation, Youtu-GraphRAG leverages a v…
Figure 7
Figure 7. Figure 7: We showcase the generalizability over six bench￾mark datasets in terms of both open and reject accuracy. Furthermore, we summarize the top-10 results in [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. MG$^2$-RAG: Multi-Granularity Graph for Multimodal Retrieval-Augmented Generation

    cs.IR 2026-04 unverdicted novelty 6.0 of 10

    MG²-RAG proposes a multi-granularity graph RAG framework that constructs hierarchical multimodal nodes via entity-driven visual grounding and performs structured retrieval, delivering SOTA results on four multimodal t...

Reference graph

Works this paper leans on

30 extracted references · 9 canonical work pages · cited by 1 Pith paper

  1. [1]

    Graphrag-bench: Challenging domain-specific reasoning for evaluating graph retrieval-augmented generation

    Yilin Xiao, Junnan Dong, Chuang Zhou, Su Dong, Qianwen Zhang, Di Yin, Xing Sun, and Xiao Huang. Graphrag-bench: Challenging domain-specific reasoning for evaluating graph retrieval-augmented generation. arXiv preprint arXiv:2506.02404, 2025

    arXiv 2025

  2. [2]

    Unifying large language models and knowledge graphs: A roadmap

    Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, and Xindong Wu. Unifying large language models and knowledge graphs: A roadmap. IEEE Transactions on Knowledge and Data Engineering, 36(7): 3580–3599, 2024

    work page 2024

  3. [3]

    Knowledge graph prompting for multi-document question answering

    Yu Wang, Nedim Lipka, Ryan A Rossi, Alexa Siu, Ruiyi Zhang, and Tyler Derr. Knowledge graph prompting for multi-document question answering. In AAAI, volume 38, pages 19206–19214, 2024

    work page 2024

  4. [4]

    Knowgpt: Knowledge graph based prompting for large language models

    Qinggang Zhang, Junnan Dong, Hao Chen, Daochen Zha, Zailiang Yu, and Xiao Huang. Knowgpt: Knowledge graph based prompting for large language models. NeurIPS, 37:6052–6080, 2024

    work page 2024

  5. [5]

    G-retriever: Retrieval-augmented generation for textual graph understanding and question answering

    Xiaoxin He, Yijun Tian, Yifei Sun, Nitesh Chawla, Thomas Laurent, Yann LeCun, Xavier Bresson, and Bryan Hooi. G-retriever: Retrieval-augmented generation for textual graph understanding and question answering. NeurIPS, 37:132876–132907, 2024

    work page 2024

  6. [6]

    Hierarchy- aware multi-hop question answering over knowledge graphs

    Junnan Dong, Qinggang Zhang, Xiao Huang, Keyu Duan, Qiaoyu Tan, and Zhimeng Jiang. Hierarchy- aware multi-hop question answering over knowledge graphs. In The Web Conf, 2023

    work page 2023

  7. [7]

    Graph retrieval-augmented generation: A survey

    Boci Peng, Yun Zhu, Yongchao Liu, Xiaohe Bo, Haizhou Shi, Chuntao Hong, Yan Zhang, and Siliang Tang. Graph retrieval-augmented generation: A survey. arXiv preprint arXiv:2408.08921, 2024

    arXiv 2024

  8. [8]

    Retrieval-augmented generation with graphs (graphrag)

    Haoyu Han, Yu Wang, Harry Shomer, Kai Guo, Jiayuan Ding, Yongjia Lei, Mahantesh Halappanavar, Ryan A Rossi, Subhabrata Mukherjee, Xianfeng Tang, et al. Retrieval-augmented generation with graphs (graphrag). arXiv preprint arXiv:2501.00309, 2024

    arXiv 2024

Show all 30 references
  1. [9]

    Modality- aware integration with large language models for knowledge-based visual question answering

    Junnan Dong, Qinggang Zhang, Huachi Zhou, Daochen Zha, Pai Zheng, and Xiao Huang. Modality- aware integration with large language models for knowledge-based visual question answering. In ACL, pages 2417–2429. ACL, 2024

    2024

  2. [10]

    From local to global: A graph rag approach to query-focused summarization

    Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, Dasha Metropolitansky, Robert Osazuwa Ness, and Jonathan Larson. From local to global: A graph rag approach to query-focused summarization. arXiv preprint arXiv:2404.16130, 2024

    2024 arXiv

  3. [11]

    Lightrag: Simple and fast retrieval- augmented generation

    Zirui Guo, Lianghao Xia, Yanhua Yu, Tu Ao, and Chao Huang. Lightrag: Simple and fast retrieval- augmented generation. arXiv preprint arXiv:2410.05779, 2024

    2024 arXiv

  4. [12]

    Gnn-rag: Graph neural retrieval for large language model reasoning

    Costas Mavromatis and George Karypis. Gnn-rag: Graph neural retrieval for large language model reasoning. arXiv preprint arXiv:2405.20139, 2024

    2024 arXiv

  5. [13]

    Gfm-rag: graph foundation model for retrieval augmented generation

    Linhao Luo, Zicheng Zhao, Gholamreza Haffari, Dinh Phung, Chen Gong, and Shirui Pan. Gfm-rag: graph foundation model for retrieval augmented generation. arXiv preprint arXiv:2502.01113, 2025

    2025

  6. [14]

    Hipporag: Neurobiolog- ically inspired long-term memory for large language models

    Bernal Jimenez Gutierrez, Yiheng Shu, Yu Gu, Michihiro Yasunaga, and Yu Su. Hipporag: Neurobiolog- ically inspired long-term memory for large language models. NeurIPS, 37:59532–59569, 2024

    2024

  7. [15]

    From rag to memory: Non-parametric continual learning for large language models

    Bernal Jiménez Gutiérrez, Yiheng Shu, Weijian Qi, Sizhe Zhou, and Yu Su. From rag to memory: Non-parametric continual learning for large language models. ICML, 2025

    2025

  8. [16]

    Raptor: Recursive abstractive processing for tree-organized retrieval

    Parth Sarthi, Salman Abdullah, Aditi Tuli, Shubh Khanna, Anna Goldie, and Christopher D Manning. Raptor: Recursive abstractive processing for tree-organized retrieval. In The Twelfth International Conference on Learning Representations, 2024

    2024

  9. [17]

    Eˆ 2graphrag: Streamlining graph-based rag for high efficiency and effectiveness

    Yibo Zhao, Jiapeng Zhu, Ye Guo, Kangkang He, and Xiang Li. Eˆ 2graphrag: Streamlining graph-based rag for high efficiency and effectiveness. arXiv preprint arXiv:2505.24226, 2025. 15 Youtu-GraphRAG

    2025 arXiv

  10. [18]

    From louvain to leiden: guaranteeing well- connected communities

    Vincent A Traag, Ludo Waltman, and Nees Jan Van Eck. From louvain to leiden: guaranteeing well- connected communities. Scientific reports, 9(1):1–12, 2019

    2019

  11. [19]

    Hotpotqa: A dataset for diverse, explainable multi-hop question answering

    Zhilin Yang, Peng Qi, Saizheng Zhang, Yoshua Bengio, William W Cohen, Ruslan Salakhutdinov, and Christopher D Manning. Hotpotqa: A dataset for diverse, explainable multi-hop question answering. arXiv preprint arXiv:1809.09600, 2018

    2018 arXiv

  12. [20]

    Musique: Multihop questions via single-hop question composition.Transactions of the Association for Computational Linguistics, 10:539–554, 2022

    Harsh Trivedi, Niranjan Balasubramanian, Tushar Khot, and Ashish Sabharwal. Musique: Multihop questions via single-hop question composition.Transactions of the Association for Computational Linguistics, 10:539–554, 2022

    2022

  13. [21]

    Constructing a multi-hop qa dataset for comprehensive evaluation of reasoning steps

    Xanh Ho, Anh-Khoa Duong Nguyen, Saku Sugawara, and Akiko Aizawa. Constructing a multi-hop qa dataset for comprehensive evaluation of reasoning steps. arXiv preprint arXiv:2011.01060, 2020

    2011 arXiv

  14. [22]

    A survey of graph retrieval-augmented generation for customized large language models

    Qinggang Zhang, Shengyuan Chen, Yuanchen Bei, Zheng Yuan, Huachi Zhou, Zijin Hong, Junnan Dong, Hao Chen, Yi Chang, and Xiao Huang. A survey of graph retrieval-augmented generation for customized large language models. arXiv preprint arXiv:2501.13958, 2025

    2025

  15. [23]

    Libo Qin, Qiguang Chen, Yuhang Zhou, Zhi Chen, Yinghui Li, Lizi Liao, Min Li, Wanxiang Che, and Philip S. Yu. Multilingual large language model: A survey of resources, taxonomy and frontiers. CoRR, abs/2404.04925, 2024. doi: 10.48550/ARXIV .2404.04925. URL https://doi.org/10.4...

    2024 doi arXiv

  16. [24]

    Natural language understanding and inference with MLLM in visual question answering: A survey

    Jiayi Kuang, Ying Shen, Jingyou Xie, Haohao Luo, Zhe Xu, Ronghao Li, Yinghui Li, Xianfeng Cheng, Xika Lin, and Yu Han. Natural language understanding and inference with MLLM in visual question answering: A survey. ACM Comput. Surv. , 57(8):190:1–190:36, 2025. doi: 10.1145/3711...

    2025 doi

  17. [25]

    Clr-bench: Evaluating large language models in college-level reasoning

    Junnan Dong, Zijin Hong, Yuanchen Bei, Feiran Huang, Xinrun Wang, and Xiao Huang. Clr-bench: Evaluating large language models in college-level reasoning. arXiv preprint arXiv:2410.17558, 2024

    2024 arXiv

  18. [26]

    Libo Qin, Qiguang Chen, Xiachong Feng, Yang Wu, Yongheng Zhang, Yinghui Li, Min Li, Wanxiang Che, and Philip S. Yu. Large language models meet NLP: A survey. CoRR, abs/2405.12819, 2024. doi: 10.48550/ARXIV .2405.12819. URLhttps://doi.org/10.48550/arXiv.2405.12819

    2024 doi arXiv

  19. [27]

    Reasoning on graphs: Faithful and interpretable large language model reasoning

    Linhao Luo, Yuan-Fang Li, Gholamreza Haffari, and Shirui Pan. Reasoning on graphs: Faithful and interpretable large language model reasoning. arXiv preprint arXiv:2310.01061, 2023

    2023 arXiv

  20. [28]

    Graphs meet ai agents: Taxonomy, progress, and future opportunities

    Yuanchen Bei, Weizhi Zhang, Siwen Wang, Weizhi Chen, Sheng Zhou, Hao Chen, Yong Li, Jiajun Bu, Shirui Pan, Yizhou Yu, et al. Graphs meet ai agents: Taxonomy, progress, and future opportunities. arXiv preprint arXiv:2506.18019, 2025

    2025 arXiv

  21. [29]

    Qa-gnn: Reasoning with language models and knowledge graphs for question answering

    Michihiro Yasunaga, Hongyu Ren, Antoine Bosselut, Percy Liang, and Jure Leskovec. Qa-gnn: Reasoning with language models and knowledge graphs for question answering. arXiv preprint arXiv:2104.06378, 2021

    2021 arXiv

  22. [30]

    Graph- constrained reasoning: Faithful reasoning on knowledge graphs with large language models

    Linhao Luo, Zicheng Zhao, Gholamreza Haffari, Yuan-Fang Li, Chen Gong, and Shirui Pan. Graph- constrained reasoning: Faithful reasoning on knowledge graphs with large language models. arXiv preprint arXiv:2410.13080, 2024. 16 Youtu-GraphRAG A Prompt templates in LLMs generatio...

    2024 arXiv

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.