A model for prejudiced learning in noisy environments

Based on the heuristics that maintaining presumptions can be beneficial in uncertain environments, we propose a set of basic requirements for learning systems to incorporate the concept of prejudice. The simplest, memoryless model of a deterministic learning rule obeying the axioms is constructed, and shown to be equivalent to the logistic map. The system's performance is analysed in an environment in which it is subject to external randomness, weighing learning defectiveness against stability gained. The corresponding random dynamical system with inhomogeneous, additive noise is studied, and shown to exhibit the phenomena of noise induced stability and stochastic bifurcations. The overall results allow for the interpretation that prejudice in uncertain environments can entail a considerable portion of stubbornness as a secondary phenomenon.