arXiv 2406.070342024-06-11cs.AIcs.CL

利用上下文感知查询表示学习改进知识图谱中的多跳逻辑推理

Improving Multi-hop Logical Reasoning in Knowledge Graphs with Context-Aware Query Representation Learning

Sungkyunkwan University（成均馆大学）；Indiana University Indianapolis（印第安纳大学印第安纳波利斯分校）

Jeonghoon Kim, Heesoo Jung, Hyeju Jang, Hogun Park

备注：: Accepted to ACL 2024 Findings

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

针对知识图谱 FOL 查询推理中几何或概率参数建模忽视查询上下文的问题，本文提出模型无关的双上下文查询表示学习方法，持续增强三个基础模型，最高提升 19.5%。

摘要：

知识图谱上的多跳逻辑推理是自然语言处理中的一项关键任务，已有大量方法旨在回答一阶逻辑（FOL）查询。近期基于几何（例如 box、cone）和概率（例如 beta 分布）的方法已经有效处理了复杂的 FOL 查询。然而，这些方法面临的一个共同挑战在于如何为这些查询确定准确的几何边界或概率参数。该挑战之所以出现，是因为现有方法依赖其计算图中的线性顺序操作，忽视了查询的逻辑结构以及可从查询关系中收集的关系诱导信息；我们将这些信息称为查询的上下文。为解决该问题，我们提出了一种模型无关的方法，通过充分融合 FOL 查询图的上下文来提升现有多跳逻辑推理方法的有效性。我们的方法独特地区分：（1）查询结构中固有的结构上下文，以及（2）查询图中每个节点在相应知识图谱中所描绘的、特有的关系诱导上下文。这种双上下文范式有助于查询图中的节点在整个多跳推理步骤中获得精炼的内部表示。通过在两个数据集上的实验，我们的方法持续增强了三个多跳推理基础模型，实现了最高 19.5% 的性能提升。我们的代码可在 https://github.com/kjh9503/caqr 获取。

英文摘要：

Multi-hop logical reasoning on knowledge graphs is a pivotal task in natural language processing, with numerous approaches aiming to answer First-Order Logic (FOL) queries. Recent geometry (e.g., box, cone) and probability (e.g., beta distribution)-based methodologies have effectively addressed complex FOL queries. However, a common challenge across these methods lies in determining accurate geometric bounds or probability parameters for these queries. The challenge arises because existing methods rely on linear sequential operations within their computation graphs, overlooking the logical structure of the query and the relation-induced information that can be gleaned from the relations of the query, which we call the context of the query. To address the problem, we propose a model-agnostic methodology that enhances the effectiveness of existing multi-hop logical reasoning approaches by fully integrating the context of the FOL query graph. Our approach distinctively discerns (1) the structural context inherent to the query structure and (2) the relation-induced context unique to each node in the query graph as delineated in the corresponding knowledge graph. This dual-context paradigm helps nodes within a query graph attain refined internal representations throughout the multi-hop reasoning steps. Through experiments on two datasets, our method consistently enhances the three multi-hop reasoning foundation models, achieving performance improvements of up to 19.5%. Our code is available at https://github.com/kjh9503/caqr.

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。

AI总结：

摘要：

英文摘要：

相关论文