Towards a model of autonomous program synthesis mechanism in the brain: Acquisition and execution mechanism of symbol processing instructions by two-layer Bayesian network

Abstract

In this paper, we first describe the concept of the architecture of an intelligent agentthat acquires knowledge from experience. Next, as one of the important elemental technologies ofthe architecture, we propose a method for pattern acquisition and pattern matching using twolayered Bayesian networks. This method uses a specially designed conditional probability model.