IndexRAG: Bridging Facts for Cross-Document Reasoning at Index Time

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bao, Zhenghua, Shi, Yi
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918393707233280
author Bao, Zhenghua
Shi, Yi
author_facet Bao, Zhenghua
Shi, Yi
contents Multi-hop question answering (QA) requires reasoning across multiple documents, yet existing retrieval-augmented generation (RAG) approaches address this either through graph-based methods requiring additional online processing or iterative multi-step reasoning. We present IndexRAG, a novel approach that shifts cross-document reasoning from online inference to offline indexing. IndexRAG identifies bridge entities shared across documents and generates bridging facts as independently retrievable units, requiring no additional training or fine-tuning. Experiments on three widely-used multi-hop QA benchmarks (HotpotQA, 2WikiMultiHopQA, MuSiQue) show that IndexRAG improves F1 over Naive RAG by 4.6 points on average, while requiring only single-pass retrieval and a single LLM call at inference time. When combined with IRCoT, IndexRAG outperforms all graph-based baselines on average, including HippoRAG and FastGraphRAG, while relying solely on flat retrieval. Our code will be released upon acceptance.
format Preprint
id arxiv_https___arxiv_org_abs_2603_16415
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle IndexRAG: Bridging Facts for Cross-Document Reasoning at Index Time
Bao, Zhenghua
Shi, Yi
Computation and Language
Artificial Intelligence
Information Retrieval
Multi-hop question answering (QA) requires reasoning across multiple documents, yet existing retrieval-augmented generation (RAG) approaches address this either through graph-based methods requiring additional online processing or iterative multi-step reasoning. We present IndexRAG, a novel approach that shifts cross-document reasoning from online inference to offline indexing. IndexRAG identifies bridge entities shared across documents and generates bridging facts as independently retrievable units, requiring no additional training or fine-tuning. Experiments on three widely-used multi-hop QA benchmarks (HotpotQA, 2WikiMultiHopQA, MuSiQue) show that IndexRAG improves F1 over Naive RAG by 4.6 points on average, while requiring only single-pass retrieval and a single LLM call at inference time. When combined with IRCoT, IndexRAG outperforms all graph-based baselines on average, including HippoRAG and FastGraphRAG, while relying solely on flat retrieval. Our code will be released upon acceptance.
title IndexRAG: Bridging Facts for Cross-Document Reasoning at Index Time
topic Computation and Language
Artificial Intelligence
Information Retrieval
url https://arxiv.org/abs/2603.16415