概 率 逻 辑 学 习 概 述

概率逻辑学习概述

陈建中¹, 刘大有²,³, 孙舒杨²,³, Stephen Muggleton¹

1. 英国帝国理工学院计算机系, 英国伦敦南肯辛顿 SW7 2AZ； 2. 吉林大学计算机科学与技术学院, 长春 130012； 3. 吉林大学符号计算与知识工程教育部重点实验室, 长春 130012

收稿日期:2006-05-18 修回日期:1900-01-01 出版日期:2006-08-26 发布日期:2006-11-26
通讯作者: 陈建中

Survey of Probabilistic Logic Learning

CHEN Jianzhong¹, LIU Dayou²,³, SUN Shuyang²,³, Stephen Muggleton¹

1. Department of Computing, Imperial College London, South Kensington SW7 2AZ, London, UK; 2. College of Computer Science and Technology, Jilin University, Changchun 130012, China; 3. Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun 130012, China

Received:2006-05-18 Revised:1900-01-01 Online:2006-08-26 Published:2006-11-26
Contact: CHEN Jianzhong

摘要/Abstract

摘要： 在归纳逻辑编程研究的基础上, 给出了概率逻辑学习的一种形式化定义框架. 根据不同概率逻辑学习方法在实例表示和概率定义上的不同, 讨论了3种概率逻辑学习框架：解释学习、证明学习和蕴涵学习; 并据此对现有的典型概率逻辑学习方法和系统进行了分析和概述.

关键词: 概率逻辑学习, 统计关系学习, 归纳逻辑编程, 学习框架

Abstract: A formal definition of Probabilistic Logic Learning is presented on the basis of Inductive Logic Programming research. Three learning settings for various Probabilistic Logic Learning approaches: probabilistic learning from interpretations, proofs and entailment are addressed based on their representation of examples and probabilistic processes. The paperinvestigates a survey of some existing Probabilistic Logic Learning approaches and systems with the three settings.

Key words: probabilistic logic learning, statistical relational learning, inductive logic programming, learning settings

中图分类号:

TP30

陈建中, 刘大有,, 孙舒杨,, Stephen Muggleton. 概率逻辑学习概述[J]. J4, 2006, 44(06): 73-79.

CHEN Jianzhong, LIU Dayou,, SUN Shuyang,, Stephen Muggleton. Survey of Probabilistic Logic Learning[J]. J4, 2006, 44(06): 73-79.

[1]	孙成敏, 刘大有, 孙舒杨. RSILP模型若干问题的研究[J]. J4, 2007, 45(02): 239-243.
[2]	孙舒杨, 刘大有, 孙成敏. 基于后验概率的Markov逻辑网参数学习方法[J]. J4, 2006, 44(06): 946-950.