首页> 外文会议>20th European conference on artificial intelligence >Maxi-Consistent Operators in Argumentation
【24h】

Maxi-Consistent Operators in Argumentation

机译:引数中的极大一致算子

获取原文
获取原文并翻译 | 示例

摘要

This paper studies an instantiation of Dung-style argumentation system with classical prepositional logic. Our goal is to explore the link between the result obtained by using argumentation to deal with an inconsistent knowledge base and the result obtained by using maximal consistent subsets of the same knowledge base. Namely, for a given attack relation and semantics, we study the question: does every extension of the argumentation system correspond to exactly one maximal consistent subset of the knowledge base? We study the class of attack relations which satisfy that condition. We show that such a relation must be conflict-dependent, must not be valid, must not be conflict-complete, must not be symmetric etc. Then, we show that some attack relations serve as lower or upper bounds with respect to the condition we study (e.g. we show that if an attack relation contains "canonical undercut" then it does not satisfy this condition). By using our results, we show for each attack relation and each semantics whether or not they satisfy the aforementioned condition. Finally, we interpret our results and discuss more general questions, like does (and when) this link is a desirable property. This work will help us obtain our long-term goal, which is to better understand the role of argumentation and, more particularly, the expressivity of logic-based instantiations of Dung-style argumentation frameworks.
机译:本文研究了具有经典介词逻辑的Dung式论证系统的实例化。我们的目标是探索通过使用论点处理不一致的知识库所得到的结果与通过使用相同知识库的最大一致子集所获得的结果之间的联系。即,对于给定的攻击关系和语义,我们研究以下问题:论证系统的每个扩展是否恰好对应于知识库的一个最大一致子集?我们研究满足该条件的攻击关系的类别。我们证明这种关系必须是冲突相关的,必须是无效的,不能是冲突完全的,不能是对称的等等。然后,我们证明某些攻击关系对于我们所处的条件起着上下限的作用。研究(例如,我们表明,如果攻击关系包含“规范底切”,则它不满足此条件)。通过使用我们的结果,我们为每种攻击关系和每种语义显示了它们是否满足上述条件。最后,我们解释我们的结果并讨论更多一般性问题,例如(何时)此链接是理想的属性。这项工作将帮助我们获得长期目标,即更好地理解论证的作用,尤其是了解Dung风格论证框架的基于逻辑的实例的表达性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号