Metabolic Cognition Engine: Toward Autonomous Cognitive Agents in Dynamic Meaning Fields

This paper introduces the Metabolic Cognition Engine (MCE), a conceptual and computational framework describing cognition as a dynamic process unfolding within evolving meaning fields. Building upon the MAD-ATT framework of metabolic attention dynamics, the model extends cognitive architecture toward autonomous agents capable of navigating informational environments through attention metabolism and semantic gradients. The framework integrates several interacting components: • Meaning fields derived from semantic embeddings• Metabolic attention dynamics governed by P · V = k · T• Multi-agent cognitive interaction• Skepticism as a structural stabilizing mechanism• Temporal evolution of informational signals• Self-organizing emergence of thematic structures Within the proposed architecture, informational signals generate gradients in a meaning field that guide the movement of cognitive agents. Agents redistribute attention across the landscape while maintaining metabolic balance between attention pressure, attention volume, and cognitive temperature. The work introduces the Metabolic Cognitive Kernel (MCK), a cyclic processing architecture in which signals, attention allocation, multi-agent interaction, memory integration, and meaning stabilization occur in continuous metabolic loops. To demonstrate feasibility, a computational prototype implemented in Python simulates cognitive agents navigating dynamic meaning landscapes derived from informational signals and semantic embeddings. The results suggest a new paradigm for artificial cognition in which intelligence emerges not from static computation but from the interaction between agents, signals, and evolving informational environments. The proposed framework contributes to ongoing research in cognitive architectures, complex adaptive systems, attention-based AI, and semantic information dynamics. Supplementary materials include simulation prototypes and conceptual diagrams illustrating the architecture of the Metabolic Cognition Engine.