QA-GraphRAG: Query-Adaptive Plug-and-Play Retrieval Integration for Graph-based Retrieval-Augmented Generation

Summary: Identifies a Local/Global query split exposing graph-RAG’s weakness on simple facts and costs from fixed retrieval. QA-GraphRAG adds a plug-and-play learned router that selects the appropriate knowledge hierarchy, improving existing GraphRAG systems. (summarized by gpt-5.6-luna on Aug 17 2026)

Paper ID: hf850e8f94a2e94c8
Venue: VLDB
Year: 2026
Pagerank: 4.9793485e-05
Overall Rank: 10,781 | 27.52%
DOI: 10.14778/3819518.3819546

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Zeang Sheng (Peking University)
2. Ruihong Sun (Tencent)
3. Jiahao Xu (Tencent)
4. Hanmei Luo (Tencent)
5. Peng Chen (Tencent)
6. Wentao Zhang (Peking University; Zhongguancun Academy)
7. Bin Cui (Peking University)

BibTeX Citation

@article{sheng_vldb26,
        title = {{QA-GraphRAG: Query-Adaptive Plug-and-Play Retrieval Integration for Graph-based Retrieval-Augmented Generation}},
        author = {Sheng, Zeang and Sun, Ruihong and Xu, Jiahao and Luo, Hanmei and Chen, Peng and Zhang, Wentao and Cui, Bin},
        journal = {PVLDB},
        series = {{VLDB} '26},
        volume = {19},
        number = {9},
        pages = {2224--2233},
        doi = {10.14778/3819518.3819546},
        url = {https://doi.org/10.14778/3819518.3819546},
        year = {2026}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 2 of 2 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
3,466	In-depth Analysis of Graph-based RAG in a Unified Framework	2025	VLDB	7.2783993e-05
4,012	LEGO-GraphRAG: Modularizing Graph-based Retrieval-Augmented Generation for Design Space Exploration	2025	VLDB	6.8550917e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	10,213	A Universal Question-Answering Platform for Knowledge Graphs	2023	SIGMOD
2	10,890	RAGPerf: An End-to-End Benchmarking Framework for Retrieval-Augmented Generation Systems	2026	VLDB
3	10,502	Skyline Retrieval meets Set-Cover Chunk Merging: A Cost-Effective RAG-Sketch for Long-Context LLM QA	2026	SIGMOD
4	10,646	From Prefix Cache to Fusion RAG Cache: Accelerating LLM Inference in Retrieval-Augmented Generation	2026	SIGMOD
5	10,963	GraphContainer: A Unified Platform for Comparing and Debugging Graph RAG Methods	2026	VLDB
6	10,791	BookRAG: A Hierarchical Structure-aware Index-based Approach for Retrieval-Augmented Generation on Complex Documents	2026	VLDB
7	10,800	MGRAG: Semantic Subgraph Matching and Graph-Aware Caching for Multimodal Retrieval-Augmented Generation	2026	VLDB
8	4,012	LEGO-GraphRAG: Modularizing Graph-based Retrieval-Augmented Generation for Design Space Exploration	2025	VLDB
9	11,029	Graph-Based Retrieval-Augmented Generation: Applications, Challenges, Solutions, and Opportunities	2026	VLDB
10	3,466	In-depth Analysis of Graph-based RAG in a Unified Framework	2025	VLDB