S-AI-DEF : A BIO-INSPIRED AND PARSIMONIOUS COGNITIVE ARCHITECTURE FOR ADAPTIVE ARTIFICIAL INTELLIGENCE IN DEFENSE SYSTEMS

In mission-critical and military environments, artificial intelligence systems must operate under strict constraints of robustness, responsiveness, and resource frugality. This paper introduces S-AI-DEF, a bio-inspired and cognitively grounded architecture derived from the Sparse Artificial Intelligence (S- AI) paradigm. It combines modular specialized agents, a central MetaAgent for symbolic orchestration, and a biologically inspired hormonal signaling layer to enable context-aware, fail-tolerant, and explainable decision-making. The architecture supports graceful degradation, local autonomy, and emergent behavior, while maintaining global coherence through top-down hormonal modulation. S-AI- DEF integrates three core cognitive principles : (1) modular specialization for taskfocused efficiency, (2) hormonal signaling for implicit coordination and urgency modulation, and (3) symbolic memory- based reasoning for strategic adaptation. This design enables robust operation under adversarial, uncertain, or degraded conditions. Experimental simulations demonstrate the system’s capacity to prioritize vital functions, self-regulate in partial failure scenarios, and provide interpretable outputs traceable to symbolic pathways. By merging cognitive modularity with biologically inspired regulation, S-AI-DEF offers a foundation for next-generation military AI : resilient by design, frugal by necessity, and controllable by principle.