arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2202.00898cs.LOcs.AI
收藏

本体论概念的量化与聚合

Quantification and Aggregation over Concepts of the Ontology

KU Leuven(鲁汶大学)

Pierre Carbonnelle, Matthias Van der Hallen, Marc Denecker

AI总结:

本文提出一阶逻辑的扩展以支持对本体概念集合的量化与聚合,区别于二阶量化,并扩展了FO(.)语言与IDP-Z3引擎,实现了无需具体化的细化容忍建模。

摘要:

我们认为,在某些知识表示(KR)应用中,我们希望量化由词汇表中的符号形式化表示的概念集合。我们表明,这种量化应区别于二阶量化和元编程量化。我们还研究了其与内涵逻辑中概念的关系。我们提出了一阶逻辑的一种扩展以支持此类抽象,并表明它允许编写具有细化容忍度的知识表达式。为避免该形式体系中出现无意义的句子,我们细化了良构句子的概念,并提出了一种验证良构性的方法,其复杂度与公式中的标记数量呈线性关系。我们相应地扩展了知识表示语言FO(.)及其推理引擎IDP-Z3。我们表明,这种扩展对于以细化容忍的方式(即无需具体化)准确建模各种问题领域至关重要。

英文摘要:

We argue that in some KR applications, we want to quantify over sets of concepts formally represented by symbols in the vocabulary. We show that this quantification should be distinguished from second-order quantification and meta-programming quantification. We also investigate the relationship with concepts in intensional logic. We present an extension of first-order logic to support such abstractions, and show that it allows writing expressions of knowledge that are elaboration tolerant. To avoid nonsensical sentences in this formalism, we refine the concept of well-formed sentences, and propose a method to verify well-formedness with a complexity that is linear with the number of tokens in the formula. We have extended FO(.), a Knowledge Representation language, and IDP-Z3, a reasoning engine for FO(.), accordingly. We show that this extension was essential in accurately modelling various problem domains in an elaboration-tolerant way, i.e., without reification.

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。

↑