摘要
统计关系学习是人工智能领域的一个新研究热点,它将关系表示、似然性理论和机器学习相结合,能更好地解决现实世界中复杂的关系数据问题,在生物信息学、Web导航、社会网、地理信息系统和自然语言理解等领域有着重要的应用.首先对统计关系学习的研究内容以及研究任务进行了介绍和总结,然后根据概率表示和推理机制的不同,对当前的统计关系学习方法进行了分类,并对各类方法进行了详细介绍,最后讨论了当前统计关系学习存在的问题,并指出了今后研究和发展的方向.
Interest in statistical relational learning （SRL） has grown rapidly in recent years. SRL integrates the relational or logical representations, probabilistic reasoning mechanisms with machine learning, and it can solve many complicated relational problems in real world. It has important applications in many fields such as World Wide Web, social networks, computational biology, information extraction, computer vision, speech recognition etc. In the past few years, SRL has received a lot of attention and a rich variety of approaches have been developed by many researchers, and they have different relational or logical representations, probabilistic reasoning mechanisms or machine learning principles. The goal of this paper is to provide an introduction to and an overview of these works. First the research fields and different tasks of SRL are introduced and summarized. And then an introductory survey and overview of the SRL approaches is provided, and the approaches are classified into four families, which are the approaches based on Bayesian networks, stochastic grammars, Markov networks, and （hidden） Markov models, according to probabilistie representations and reasoning mechanisms. For each approach family, the probabilistic logical models, parameter estimation and structure learning, and the states-of-the-art are introduced. Finally, the current problems in SRL are discussed and future research directions are pointed out.
出处
《计算机研究与发展》
EI
CSCD
北大核心
2008年第12期2110-2119,共10页
Journal of Computer Research and Development
基金
国家自然科学基金重大项目(60496321)
国家自然科学基金项目(60573073
60503016
60603030
60773099
60703022)
国家"八六三"高技术研究发展计划基金项目(2006AA10Z245
2006AA10A309)
吉林省科技发展计划重点项目(20060213)
吉林省科技发展计划基金项目(20030523)
欧盟项目TH/Asia Link/010(111084)~~
关键词
统计关系学习
似然逻辑学习
多关系数据挖掘
统计学习
关系学习
statistical relational learning
probabilistic logic learning
multi-relational data mining
statistical learning
relational learning
