基于贝叶斯网络的概率图模型变分近似推理研究

Variational Approximate Reasoning for Probabilistic Graph Models Based on Bayesian Networks

概率图模型利用了概率论和图论的关键内容,为解决多变量关系中所存在的依赖性与复杂性提供了完整的解决路径。在语言处理、计算机视觉、计算生物学等众多领域得到广泛应用。本文以贝叶斯网络作为核心研究方向,就概率图模型变分近似推理的全局性、收敛性、单调性等诸多特征进行辩证分析。最终论述了基于贝叶斯网络的概率图模型变分近似推理算法的应用范畴和优势,以便为相关研究提供理论参考。

Probabilistic graph models use the key content of probability theory and graph theory to provide a complete solution to the dependence and complexity existing in multivariate relations.It has been widely used in many fields such as language processing,computer vision, computational biology and so on.This paper takes the Bayesian network as the core research direction,and conducts a dialectical analysis on the global,convergence and monotonic characteristics of the variational approximate reasoning of the probability map model.Finally,the application category and advantages of variational approximate reasoning algorithm based on Bayesian network probability model are discussed in order to provide theoretical reference for related research.