发表机构
Institute of Computational Linguistics, Heidelberg University; Aleph Alpha Research（海德堡大学计算语言学研究所; Aleph Alpha 研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出GLM-RAG的GLM基检索器，对比三类检索器性能，发现微调GLM检索器跨域泛化更优且在多跳基准达SOTA，GNN与向量搜索各有优势。
AI 中文摘要
知识图谱上的检索增强生成（RAG）需要能有效捕捉图结构和语义信息的检索器。近期研究探索了基于图神经网络（GNN）的检索器，用于在多跳推理任务中建模图拓扑；同时，图语言模型（GLM）作为融合图推理与语言模型语义能力的有前景范式已出现。本研究提出一种基于GLM的检索器，在单跳、多跳RAG设置中对比GLM基、GNN基及传统向量搜索基检索器的相对优势，尤其关注其对未见领域的迁移能力。研究发现，微调后的GLM检索器跨领域泛化能力更优，在两个多跳基准上实现SOTA；在领域内多跳问答数据集上，其表现与现有方法相当，且随参数和子图覆盖度提升呈现良好的可扩展性。GNN基检索器训练设置高效，能实现更高的图覆盖度；而向量搜索基线在单跳数据集上表现更出色。
英文摘要
Retrieval-augmented generation (RAG) over knowledge graphs requires retrievers that can effectively capture both graph structure and semantic information. Recent approaches have explored graph neural network (GNN)-based retrievers to model graph topology in multi-hop reasoning tasks. In parallel, graph language models (GLMs) have emerged as a promising paradigm that integrates graph reasoning and the semantic capabilities of language models. In this work, we introduce a GLM-based retriever and investigate the comparative strengths of GLM-based, GNN-based, and traditional vector-search-based retrievers in single- and multi-hop RAG settings, and with a particular focus on transferability to unseen domains. Our findings suggest that finetuned GLM retrievers generalize better out of domain, achieving SOTA on two multi-hop benchmarks. On in-domain multi-hop QA datasets they remain comparable to prior work, with promising scaling as parameters and subgraph coverage increase. GNN-based retrievers achieve higher graph coverage with an efficient training setup, whereas the vector-search baseline excels at single-hop datasets.
CommentsAccepted to AACL-IJCNLP 2026, 9 pages, 19 figures
