This work presents a conceptual, object-centered architecture for emotional evaluation and attention allocation in artificial agents. The proposed approach treats emotional systems as computational mechanisms for assessing the subjective significance of objects and prioritizing computational resources in dynamic environments. The central concept of the architecture is the emotional profile of an object, representing predicted affective responses prior to direct interaction. The model introduces a hierarchical attention mechanism driven by physical constraints, emotional object evaluation, and cognitive focusing. The work is theoretical in nature and is intended as a foundation for future empirical research and computational implementations in emotionally grounded artificial intelligence.
Assylkhan Amrekulov (Fri,) studied this question.