Modelling Intelligent Systems Using COH: A Unified Framework for Intelligence and Intelligent Systems

Constrained Object Hierarchies (COH) present a neuroscience-grounded theoretical framework for artificial general intelligence (AGI) that enables unified modelling of intelligent systems across diverse application domains. This paper introduces the COH framework as a comprehensive 9-tuple formalism that captures the essential components of intelligent systems through compositional hierarchy, adaptive neural components, and constraint-based reasoning. We demonstrate the framework's versatility by modelling 15 distinct intelligent systems across domains including autonomous vehicles, healthcare, smart grids, personalized education, and cybersecurity. Each model illustrates how COH's innovative concepts—particularly identity constraints, goal constraints, trigger constraints, and constraint daemons—provide a common language for describing intelligent behavior. The paper establishes COH as a foundational theory for AGI system design, offering methodological guidance for analysis, design, and implementation of intelligent systems with provable properties and predictable behavior.