“The logical categories of learning and communication”: reconsidered from a polycontextural point of view: Learning in machines and living systems

摘要: Purpose – Bateson's model of classifying different types learning will be analyzed from a logical and technical point view. While 0 has been realized for chess playing computers, I turns out today as the basic concept artificial neural nets (ANN). All models ANN are basically (non linear) data filters, which is idea behind simple behavioristic input‐output models.Design/methodology/approach The paper discuss systems designed on it demonstrate that these do not have an environment, i.e. they non‐cognitive therefore “non‐learning” systems.Findings Models based category Learning II differ fundamentally I. They cannot modeled any longer basis classical (mono‐contextural) logics. Technical artifacts belong to this able change their algorithms (behavior) by own effort. process b...