Short Papers of the 16th International Conference on Inductive Logic Programming pages:173-175
International Conference on Inductive Logic Programming edition:16 location:Santiago de Compostela, Spain date:August 24-27, 2006
Recently, there has been an increasing interest in probabilistic logical models and a variety of such languages has been proposed. Although many authors provide high-level arguments to show that in principle models in their language can be learned from data, most such learning algorithms have not yet been studied in detail.We propose a representation where logical probability trees are used as conditional probability distributions and propose a new algorithm to learn both structure and parameters of Logical Bayesian Networks. We present experiments on relational reinforcement learning domains.