Emotional Intelligence: Abstract Cognition Innovation in Artificial General Intelligence Systems

This study investigates the incorporation of abstract emotion-triggering mechanisms into artificial general intelligence (AGI) systems through the nonaxiomatic reasoning system (NARS) framework. Leveraging cognitive appraisal theory, the proposed model facilitates dynamic regulation of cognitive resources by modulating priority and durability based on goal alignment and temporal evaluation. Distinct from conventional emotion models that depend on predefined feedback mechanisms, this framework enables generalized emotional responses, thereby enhancing adaptability to complex temporal and causal dynamics. Experimental validation conducted on flappy bird and airplane combat simulation platforms illustrates the superiority of the emotion-driven NARS, which demonstrates enhanced decision-making efficiency, robust goal prioritization, and superior adaptability compared to its nonemotional counterpart. These findings underscore the potential of emotion-enabled AGI systems to advance applications in high-stakes domains, including autonomous driving and robotics, where real-time adaptability and efficient resource allocation are paramount.