Representing Statistical Information and Degrees of Belief in First-Order Probabilistic Conditional Logic
Representing Statistical Information and Degrees of Belief in First-Order Probabilistic Conditional Logic
复制标题
表示一阶概率条件逻辑中的统计信息和置信度
DOI：
--
复制
发表时间：
2009
期刊：
影响因子：
--
通讯作者：
Matthias Thimm
中科院分区：
文献类型：
--
作者：
Matthias Thimm
将最大熵方法应用于概率推理已被证明是一种有用的常识推理方法。然而，这种逻辑和类似的形式主义的表达能力是有限的，由于它们的基础上的命题逻辑，在过去的几年中，已经提出了很多建议，在关系设置的概率推理。这些建议中的大多数依赖于传统的基于图的概率模型（如贝叶斯网或马尔可夫网）的扩展，而概率条件逻辑并不预先假定要表示的模型的任何图形结构。在本文中，我们采取的方法，提升最大熵方法的关系的情况下，使用一阶版本的概率条件逻辑。此外，我们采取了一个特殊的重点表示关系概率知识的dierentiating之间的关系概率条件的不同的直觉，即统计解释和解释之间的程度的信念。我们开发了一个列表的理想属性的推理过程，支持这些dierent解释，并提出了一个specic推理过程，fullls这些属性。最后讨论了相关的工作，并对今后的研究方向提出了一些建议.
Employing maximum entropy methods on probabilistic con- ditional logic has proven to be a useful approach for commonsense rea- soning. Yet, the expressive power of this logic and similar formalisms is limited due to their foundations on propositional logic and in the past few years a lot of proposals have been made for probabilistic reasoning in re- lational settings. Most of these proposals rely on extensions of traditional graph-based probabilistic models like Bayes nets or Markov nets whereas probabilistic conditional logic does not presuppose any graphical struc- ture underlying the model to be represented. In this paper we take an approach of lifting maximum entropy methods to the relational case by using a rst-order version of probabilistic conditional logic. Furthermore, we take a specic focus on representing relational probabilistic knowledge by dierentiating between dierent intuitions on relational probabilistic conditionals, namely between statistical interpretations and interpreta- tions on degrees of belief. We develop a list of desirable properties on an inference procedure that supports these dierent interpretations and propose a specic inference procedure that fullls these properties. We furthermore discuss related work and give some hints on future research.