Learning Low Inference Complexity Bayesian Networks

One of the main research topics in machine learning nowadays is the improvement of the inference and learning processes in probabilistic graphical models. Traditionally, inference and learning have been treated separately, but given that the structure of the model conditions the inference complexity, most learning methods will sometimes produce very inefficient inference models. In this paper we propose a new model for representing discrete probability distributions that allows efficiently evaluating the inference complexity and a framework for exact inference. We also present the procedures required to learn these models from data. Experimental results show that the new model produces accurate answers with a tractable computational cost.