The work investigates online learning, robustness under uncertainty, and autonomous regulation of behavior in dynamic environments. Computational models integrate neural adaptation mechanisms with machine learning approaches to enable flexible perception, decision-making, and action. Particular emphasis is placed on embodied intelligence, human-aware robotics, and real-time adaptation to user cognitive states. By bridging neuroscience and AI, the research advances systems capable of continuous learning, resilient performance, and context-sensitive responses. These developments contribute to next-generation intelligent agents designed for collaborative, assistive, and socially embedded applications.