Markov逻辑网研究综述

Abstract

Markov逻辑网是将Markov网络与一阶谓词逻辑相结合的统计关系学习模型。Markov逻辑网在实体识别、数据融合、信息抽取等领域都有重要研究价值，具有广泛的应用。本文较为全面的介绍了Markov逻辑网的理论模型、推理、参数学习、与其他算法的比较，最后探讨Markov逻辑网未来的研究方向。 Markov logic networks (MLNs) is a kind of statistical relational learning model which combines Markov network and first-order logic together. MLNs has the significant research value and in many areas it has widely applications, such as entity recognition, data integration and information extrac-tion. In this paper, we introduced the theoretical model of Markov logic networks, inference and pa-rametric learning of it and compared it with other. In the end, we discussed future works of MLNs.