Uncertainty and Time in Modern Artificial Intelligence

Abstract

After finding out the main reason for decision-makers distrust in computer support (low model interpretability due to low NL- capability), the paper outlines the shift towards semiotic-oriented software proposing Cybernetic Modelling, a modelling sub-species aimed at enhancing large- scale living systems model efficiency and user acceptance. Uncertainty due to future events is treated via third IF value (“Undecidable in the time span given”) avoiding probabilities. Time extension allows distinct temporal dimensions. (A Proof-of-Concept example describing a model for homeostasis with hysteretic delay is given in a related paper.)