发表机构
University of Calabria; DeMaCS（卡拉布里亚大学; DeMaCS机构）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对知识图谱与大语言模型融合的挑战，提出GRALAN模型，通过关系令牌实现KG在LLM语义空间的适配，在问答任务中显著优于现有方法，为KG-LLM融合提供新范式。
AI 中文摘要
大语言模型（LLMs）擅长推理，但需要知识图谱（KGs）提供的基础支撑。然而，将这两种范式结合颇具挑战。我们提出GRALAN，它能通过保留图结构的关系令牌让KGs直接在LLM的语义空间中“说话”。GRALAN可训练的语言中介能为任何冻结的LLM生成结构化令牌，为知识密集型应用奠定基础。我们通过将问答任务重新表述为针对问题聚焦子图的实体分类，验证了其有效性。实验表明，GRALAN的表现显著优于现有方法，尤其在复杂多跳推理任务上，它建立了一种新的KG-LLM融合范式，在保持结构保真度的同时利用了LLMs的推理能力。
英文摘要
Large Language Models (LLMs) excel at reasoning but benefit from grounding provided by Knowledge Graphs (KGs). However, integrating these paradigms is challenging. We introduce GRALAN, which enables KGs to speak directly in the LLM's semantic space through relational tokens that preserve graph structure. GRALAN-s trainable language mediator generates structured tokens for any frozen LLM, creating a foundation for knowledge-intensive applications. We demonstrate its effectiveness in question-answering by re-framing the task as entity classification over question-focused subgraphs. Experiments show that GRALAN significantly outperforms existing methods, particularly on complex multi-hop reasoning tasks, establishing a new paradigm for KG-LLM integration that maintains structural fidelity while leveraging LLMs' reasoning capabilities.
CommentsAccepted to ISWC 2025
DOI：10.1007/978-3-032-09527-5_3
