A local machine learning task in algebraical Bayesian networks: a probabilistic-logic approach

One of the problems that slow down intelligent information systems development and industry- wide spread is so-called knowledge bottleneck. Machine learning for various uncertain knowledge representations and models used in intelligent systems is a promising way to cope with the bottleneck. Algebraical Bayesian networks are a probabilistic graphical model that allows for representing and processing interval estimates of probabilities. The paper goal is to describe the machine learning task in regard to a knowledge pattern of an algebraical Bayesian network as well as to present a few ways to solve the task and to outline obstacles related to those ways.