1 graph will be a tree. This is an important distinction, as message passing algorithms are usually exact for trees, but only approximate for graphs with cycles.
2 A popular message passing algorithm is the sum-product algorithm, which efficiently computes all the marginals of the individual variables in the joint distribution
3 Markov networks and Bayesian networks can be represented as factor graphs
今天看到维基百科里面关于因子图的介绍,比较漂亮的几句话,摘录下来,以后有机会写个总结性的博客。不过先说一下大致的意思:
在概率图模型的推理过程中,有精确的和近似的两种求解算法,分别适用于不同的情况。精确推理算法只能用在简单图(无环),也就是树状图模型的推理当中。
概率图模型中一种比较基本的推理算法就是边缘化算法(sum-production),也叫变量消元法。
最后一句是说,因子图(也叫因式图)的两种特例:贝叶斯网络和马尔科夫场。