Enhancing Document-Level Question Answering via Multi-Hop Retrieval-Augmented Generation with LLaMA 3

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Xinyue, Lin, Ziqi, Sun, Fang, Zhang, Wenchao, Tong, Kejian, Liu, Yunbo
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915350725001216
author Huang, Xinyue
Lin, Ziqi
Sun, Fang
Zhang, Wenchao
Tong, Kejian
Liu, Yunbo
author_facet Huang, Xinyue
Lin, Ziqi
Sun, Fang
Zhang, Wenchao
Tong, Kejian
Liu, Yunbo
contents This paper presents a novel Retrieval-Augmented Generation (RAG) framework tailored for complex question answering tasks, addressing challenges in multi-hop reasoning and contextual understanding across lengthy documents. Built upon LLaMA 3, the framework integrates a dense retrieval module with advanced context fusion and multi-hop reasoning mechanisms, enabling more accurate and coherent response generation. A joint optimization strategy combining retrieval likelihood and generation cross-entropy improves the model's robustness and adaptability. Experimental results show that the proposed system outperforms existing retrieval-augmented and generative baselines, confirming its effectiveness in delivering precise, contextually grounded answers.
format Preprint
id arxiv_https___arxiv_org_abs_2506_16037
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing Document-Level Question Answering via Multi-Hop Retrieval-Augmented Generation with LLaMA 3
Huang, Xinyue
Lin, Ziqi
Sun, Fang
Zhang, Wenchao
Tong, Kejian
Liu, Yunbo
Computation and Language
Machine Learning
This paper presents a novel Retrieval-Augmented Generation (RAG) framework tailored for complex question answering tasks, addressing challenges in multi-hop reasoning and contextual understanding across lengthy documents. Built upon LLaMA 3, the framework integrates a dense retrieval module with advanced context fusion and multi-hop reasoning mechanisms, enabling more accurate and coherent response generation. A joint optimization strategy combining retrieval likelihood and generation cross-entropy improves the model's robustness and adaptability. Experimental results show that the proposed system outperforms existing retrieval-augmented and generative baselines, confirming its effectiveness in delivering precise, contextually grounded answers.
title Enhancing Document-Level Question Answering via Multi-Hop Retrieval-Augmented Generation with LLaMA 3
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2506.16037