Importance score

The importance score is set during enrichment and reflects how consequential an experience event is expected to be. It is not a simple function of outcome magnitude - an event can have a large immediate effect but low importance if it is routine and well-understood, or a small immediate effect but high importance if it touches a high-value goal or occurs in a novel context. The enrichment worker computes importance from a combination of outcome magnitude, goal relevance (how directly does this event relate to active goals?), novelty (how different is this from previously indexed experience?), and a configurable domain weight that operators can use to mark certain event classes as inherently more important. Importance score then flows through the rest of the pipeline as a primary weighting signal: the consolidation worker prioritizes high-importance events for semantic abstraction; retrieval priority weighting gives importance an 18% share of the final reinforcement score; and the forgetting system uses low importance score as the primary criterion for decay rate assignment, letting low-importance memories fade faster than high-importance ones.