OpenPsi brings motivation, emotion, and decision-making together in one cognitive architecture built on OpenCog Hyperon.
Inspired by Dietrich Dorner's PSI theory, OpenPsi explores how intelligent agents can develop behavior based on their internal needs and the world around them. Instead of treating emotions as separate features, the system allows emotional states to emerge naturally from the agent's goals, experiences, and decisions.
Our work focuses on turning these ideas into a practical and flexible framework for building more adaptive, human-like AI systems.
How OpenPsi works
Modulators
Modulators describe the agent's current internal state and influence how it reacts, makes decisions, and expresses emotional behavior.
Demands
Demands represent what the agent currently needs. They act as the driving force behind its behavior.
Goals
Goals are created to help the agent respond to its demands and move toward a desired outcome.
Cognitive schemas
Cognitive schemas connect situations, actions, and goals. They help the agent learn which actions are most useful in different contexts and make better decisions over time.