arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2605.25985cs.AI
收藏

面向多自由变量复杂逻辑查询的神经可扩展符号搜索框架

Neural Scalable Symbolic Search Framework for Complex Logical Queries with Multiple Free Variables

Department of Mathematical Sciences, Tsinghua University(清华大学数学科学系);Squarepoint Capital(Squarepoint资本);Department of Computer Science and Engineering, Hong Kong University of Science and Technology(香港科学与技术大学计算机科学与工程系);Department of Computer Sciences, University of Rochester(罗切斯特大学计算机科学系)

Weizhi Fei, Hang Yin, Zihao Wang, Shukai Zhao, Wei Zhang, Yangqiu Song

AI总结:

针对知识图谱上多自由变量复杂查询的联合排序难题,提出神经可扩展符号搜索(NS3)框架,通过预算约束和超节点合并近似联合排序,显著提升性能。

摘要:

复杂查询回答(CQA)是在不完整知识图谱(KG)上进行知识表示和推理的基本任务。回答带有$k$个自由变量的存在性一阶查询(即$ ext{EFO}_k$查询)是一个关键但具有挑战性的问题,因为它需要对$\mathcal{E}^k$中的答案元组进行排序,其中$\mathcal{E}$表示KG的实体集。随着$k$的增长,这很快变得难以处理。因此,现有基准和方法依赖于单个变量的边际排序;然而,边际排序是元组真实联合排序的较差代理。基于$ ext{EFO}_1$查询的神经符号搜索,我们提出了神经可扩展符号搜索(NS3),这是一个预算框架,无需枚举$\mathcal{E}^k$即可近似联合排序。NS3 (i) 回答边际化子查询以获得必要的候选集,(ii) 将多个自由变量合并为超节点,其域由动态预算$B$修剪和控制,以及(iii) 逐步将$ ext{EFO}_k$查询简化为在预算缩减域上的$ ext{EFO}_{k-1}$查询。在三个标准KG数据集上,NS3在保持强边际准确性的同时,显著提高了联合排序性能。我们进一步发布了一个联合排序基准,将现有的$ ext{EFO}_1$数据集扩展到$k=3$,从而能够系统评估多变量查询。我们的代码提供在https://github.com/HKUST-KnowComp/NS3_KDD2026。

英文摘要:

Complex Query Answering (CQA) is a fundamental knowledge representation and reasoning task over incomplete knowledge graphs (KGs). Answering existential first-order queries with $k$ free variables (i.e., $\text{EFO}_k$ queries) is a crucial yet challenging problem, as it requires ranking answer tuples in $\mathcal{E}^k$, where $\mathcal{E}$ denotes the entity set of a KG. This quickly becomes intractable as $k$ grows. Consequently, existing benchmarks and methods rely on marginal rankings over individual variables; however, marginal rankings are a poor proxy for the true joint ranking of tuples. Building on neural symbolic search for $\text{EFO}_1$ queries, we propose Neural Scalable Symbolic Search (NS3), a budgeted framework that approximates joint ranking without enumerating $\mathcal{E}^k$. NS3 (i) answers marginalized sub-queries to obtain necessary candidate sets, (ii) merges multiple free variables into hypernodes whose domains are pruned and controlled by a dynamic budget $B$, and (iii) progressively reduces an $\text{EFO}_k$ query to an $\text{EFO}_{k-1}$ query over a budgeted reduced domain. Across three standard KG datasets, NS3 substantially improves joint ranking performance while retaining strong marginal accuracy. We further release a joint-ranking benchmark that extends existing $\text{EFO}_1$ datasets to $k=3$, enabling systematic evaluation of multi-variable queries. Our code is provided in https://github.com/HKUST-KnowComp/NS3_KDD2026.

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。

↑