arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.14766cs.AIcs.LG
收藏

GR-Agent: 在不完整知识图谱下适应性图推理代理

GR-Agent: Adaptive Graph Reasoning Agent under Incomplete Knowledge

University of Oslo(奥斯陆大学);Bosch Center for AI(博世人工智能中心);University of Stuttgart(斯图加特大学);University of Oxford(牛津大学);Amazon(亚马逊);The University of Manchester(曼彻斯特大学);University of Southampton(南安普顿大学)

Dongzhuoran Zhou, Yuqicheng Zhu, Xiaxia Wang, Hongkuan Zhou, Jiaoyan Chen, Steffen Staab, Yuan He, Evgeny Kharlamov

AI总结:

GR-Agent通过构建交互环境并利用图推理工具,在不完整知识图谱中实现更有效的推理,优于现有方法。

摘要:

大型语言模型(LLMs)在知识图谱问答(KGQA)任务中取得了优异成果,但大多数基准测试假设知识图谱(KGs)是完整的,其中存在直接支持三元组。这将评估简化为浅层检索,并忽视了现实中不完整KGs的情况,其中许多事实缺失,答案必须从现有事实中推断。我们通过提出一种在KG不完整条件下构建基准的方法来填补这一空白,该方法移除了直接支持三元组,同时确保所需的替代推理路径仍可用于推断答案。在使用我们方法构建的基准测试中,现有方法在不完整情况下表现出一致的性能下降,突显了其有限的推理能力。为克服这一限制,我们提出了适应性图推理代理(GR-Agent)。它首先从KG中构建一个交互环境,然后在该环境中将KGQA形式化为代理与环境的交互。GR-Agent在由图推理工具组成的动作空间上运行,并维护一个潜在支持推理证据的记忆,包括相关关系和推理路径。广泛实验表明,GR-Agent在非训练基线中表现优于,并在完整和不完整设置下与训练基方法表现相当。

英文摘要:

Large language models (LLMs) achieve strong results on knowledge graph question answering (KGQA), but most benchmarks assume complete knowledge graphs (KGs) where direct supporting triples exist. This reduces evaluation to shallow retrieval and overlooks the reality of incomplete KGs, where many facts are missing and answers must be inferred from existing facts. We bridge this gap by proposing a methodology for constructing benchmarks under KG incompleteness, which removes direct supporting triples while ensuring that alternative reasoning paths required to infer the answer remain. Experiments on benchmarks constructed using our methodology show that existing methods suffer consistent performance degradation under incompleteness, highlighting their limited reasoning ability. To overcome this limitation, we present the Adaptive Graph Reasoning Agent (GR-Agent). It first constructs an interactive environment from the KG, and then formalizes KGQA as agent environment interaction within this environment. GR-Agent operates over an action space comprising graph reasoning tools and maintains a memory of potential supporting reasoning evidence, including relevant relations and reasoning paths. Extensive experiments demonstrate that GR-Agent outperforms non-training baselines and performs comparably to training-based methods under both complete and incomplete settings.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑