Arithmetic circuits of the noisy-or models
Abstract
In this paper we have used the size of an arithmetic circuit to show that preprocessing of a Bayesian network composed of noisy-or gates using tensor rank-one decomposition gives better results than the standard technique called parent divorcing.