# GraphER: An Efficient Graph-Based Enrichment and Reranking Method for Retrieval-Augmented Generation

Source: [GraphER: An Efficient Graph-Based Enrichment and Reranking Method for Retrieval-Augmented Generation](<https://www.sota2.com/research/paper/69c60eed0636e5bf4afc83e0>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-08-05T05:00:12.726Z

Authors: Ruizhong Miao, Yuying Wang, Rongguang Wang, Chenyang Li, Tao Sheng, Sujith Ravi, Dan Roth

Published: 2026-03-26T01:30:22.000Z

## Abstract

Semantic search in retrieval-augmented generation (RAG) systems is often insufficient for complex information needs, particularly when relevant evidence is scattered across multiple sources, because it may fail to retrieve the complete set of evidence. Existing approaches to addressing this problem either rely on iterative agentic retrieval, which can be computationally inefficient, or maintain additional structures such as knowledge graphs, which introduce storage and maintenance overhead. In this paper, we propose GraphER, a graph-based enrichment and reranking framework that (1) leverages the organizational structure of data to capture proximity relationships beyond semantic similarity, (2) constructs a graph at query time based on these proximities, and (3) applies graph-based ranking to surface the top candidate documents. Experiments across table retrieval, multi-hop retrieval, and long-document retrieval benchmarks demonstrate consistent improvements in terms of retrieval completeness. Additionally, GraphER requires no additional graph infrastructure and integrates seamlessly with standard vector stores. The framework is retriever-agnostic, supports multiple forms of proximity, and introduces minimal query-time latency.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Multi-hop QA Retrieval on HotpotQA](<https://www.sota2.com/research/sota/multi-hop-qa-retrieval-on-hotpotqa.md>): Multi-hop QA Retrieval; HotpotQA; Precision@5 (%): 78.8.

- [Multi-hop QA Retrieval on 2WikiMultihopQA](<https://www.sota2.com/research/sota/multi-hop-qa-retrieval-on-2wikimultihopqa.md>): Multi-hop QA Retrieval; 2WikiMultihopQA; PR@5: 45.7.

- [Multi-hop QA Retrieval on MuSiQue (PR@5, PR@10, QA Acc.)](<https://www.sota2.com/research/sota/multi-hop-qa-retrieval-on-musique-pr-at-5-pr-at-10-qa-acc.md>): Multi-hop QA Retrieval; MuSiQue; PR@5: 25.6.

- [Table Retrieval on Spider 1 (dev)](<https://www.sota2.com/research/sota/table-retrieval-on-spider-1-dev.md>): Table Retrieval; Spider 1 (dev); PR@5: 85.2.

- [Table Retrieval on Spider 1 (test)](<https://www.sota2.com/research/sota/table-retrieval-on-spider-1-test.md>): Table Retrieval; Spider 1 (test); Precision@5: 65.7.

- [Table Retrieval on Bird (dev)](<https://www.sota2.com/research/sota/table-retrieval-on-bird-dev.md>): Table Retrieval; BIRD (dev); Precision@5: 86.2.

- [Table Retrieval on Beaver (dev)](<https://www.sota2.com/research/sota/table-retrieval-on-beaver-dev.md>): Table Retrieval; Beaver (dev); PR@5: 17.

- [Document Retrieval on BEIR-NQ chunked documents entire query set](<https://www.sota2.com/research/sota/document-retrieval-on-beir-nq-chunked-documents-entire-query-set.md>): Document Retrieval; BEIR-NQ chunked documents entire query set; PR@5: 78.2.

- [Document Retrieval on BEIR-NQ chunked documents - queries with more than one relevant object](<https://www.sota2.com/research/sota/document-retrieval-on-beir-nq-chunked-documents-queries-with-more-than-one-relevant-object.md>): Document Retrieval; BEIR-NQ chunked documents - queries with more than one relevant object; Precision@5: 54.4.

- [Table Retrieval on Spider 1 (train)](<https://www.sota2.com/research/sota/table-retrieval-on-spider-1-train.md>): Table Retrieval; Spider 1 (train); Precision@5: 55.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.24925>)

[PDF](<https://arxiv.org/pdf/2603.24925.pdf>)