统计关系学习模型Markov逻辑网综述-计算机应用研究.PDFVIP

2007 年第 2 期 孙舒杨等: 统计关系学习模型 Markov 逻辑网综述 ·1 · * 统计关系学习模型 Markov 逻辑网综述 孙舒杨1, 2 , 刘大有1, 2 , 孙成敏1 , 2 , 黄冠利3 ( 1. 吉林大学 计算机科学与技术学院, 吉林 长春 130012; 2. 吉林大学 符号计算与知识工程教育部重点实验室, 吉林 长春 130012; 3. 北京电子科技职业学院, 北京 100029) 摘 要: 统计关系学习是人工智能研究的热点, 在生物信息学、地理信息系统和自然语言理解等领域有着重要 应用 , Markov 逻辑网是将 Markov 网与一阶逻辑相结合的一种全新的统计关系学习模型。介绍了 Markov 逻辑网 的理论模型和学习方法, 并探讨了目前存在的问题和研究方向。 关键词: 统计关系学习; 一阶逻辑; Markov 网; 机器学习; Markov 逻辑网 中图法分类号: TP30 文献标识码: A 文章编号: 1001- 3695( 2007) 02- 0001- 03 Survey of Markov Logic Networks 1, 2 1, 2 1, 2 3 SUN Shu-yang , LIU Da-you , SUN Cheng-min , HUANG Guan-li ( 1. College of Computer Science Technology, Jilin University, Changchun Jiling 130012, China; 2. Key Laboratory of Symbolic Computation Knowledge Engineering for Ministry of Education, Jilin University, Changchun Jiling 130012, China; 3 . Beijing Vocational College of Elec- tronic Science, Beijing 100029, China) Abstract: Statistical Relational Learning( SRL) is a highlight in AI research field, and has important applications on many areas, such as bioinformatics, geography information systems and natural language processing, etc. Markov Logic Networks ( MLNs) are a kind of SRL model combining Markov networks and the first-order logic together. In this paper, we introduced the theory and learning methods of MLNs, as well as discussed current problems and directions for future work. Key words: Statistical Relational Learning; First-Order Logic; Markov Networks; Machine Learning; Markov Logic Networks 近年来, 统计关系学习 ( Statistical R