AGI System Architecture Based on Hybrid Structure (RB×NN) for SUNIE and LUMIE: A Conceptual Pathway Toward Achieving Artificial General Intelligence

This paper presents a unified AGI system architecture based on a hybrid structure combining rule‑based mechanisms (RB) and neural networks (NN). The proposed architecture serves as the conceptual foundation for SUNIE and LUMIE, two next‑generation AI systems designed to achieve controllable, interpretable, and generalizable intelligence. By integrating deterministic reasoning modules with adaptive learning components, the system enables stable behavior, transparent internal processes, and scalable cognitive functions. This work outlines the core principles, structural layers, functional pathways, and theoretical rationale behind the hybrid RB×NN approach, offering a practical and conceptual roadmap toward achieving Artificial General Intelligence.