Graph-based Approaches and Functionalities in Retrieval-Augmented Generation: A Comprehensive Survey
Zulun Zhu, Tiancheng Huang, Kai Wang, Junda Ye, Xinghe Chen, Siqiang Luo · 2025
Read Graph-based Approaches and Functionalities in Retrieval-Augmented Generation: A Comprehensive Survey by Zulun Zhu, Tiancheng Huang, Kai Wang, Junda Ye, Xinghe Chen, Siqiang Luo (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-10499.txt.
Description
Zulun Zhu, Tiancheng Huang, Kai Wang, Junda Ye, Xinghe Chen, and Siqiang Luo survey how graphs power retrieval-augmented generation. Their paper from twenty twenty-five starts from a known flaw: language models hallucinate when training data is thin or stale. Retrieval systems fetch outside facts to ground replies. Much of that outside knowledge is structured, with entities tied in webs.
Graph methods tap those ties for deeper reasoning than flat text search allows. The survey sorts graph roles across the stack: building the store, the search math, the pipeline, and the end tasks. No unified review of these roles had been made. The authors map what methods share and where they part. Open challenges and next steps close the piece. It serves graph learning, database, and language readers alike.
Catalog record
| CREATOR | Zulun Zhu, Tiancheng Huang, Kai Wang, Junda Ye, Xinghe Chen, Siqiang Luo |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 137 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-10499