Belief propagation helps compute probabilities in a Bayesian network. 信念传播可以帮助在贝叶斯网络中计算概率。
In large factor graphs, belief propagation iteratively passes messages between variables and factors to approximate marginal distributions, even when exact inference is infeasible. 在大型因子图中,信念传播通过在变量与因子之间迭代传递消息来近似边缘分布,即使精确推断不可行也能发挥作用。