Glossary
Definitions of the concepts behind taste infrastructure and AI-native personalization.
Affinity
Affinity is the modeled strength of fit between a person and an item, derived from perception before responsiveness, so systems can rank by what will resonate rather than by popularity or recency.
Cold-start personalization
Cold-start personalization is the ability to model a person's taste from little or no prior interaction data, working in the cold and sparse regime where most recommender systems fail.
Resonance vs relevance
Relevance measures whether a result matches the query; resonance measures whether it matches the person's taste, which is the difference between a technically correct result and one that feels right.
Taste Graph
The Taste Graph is a structured representation of taste and affinity that makes human taste computable, so AI systems can reason about what a person will find resonant rather than merely relevant.
Taste and preference modeling
Taste and preference modeling turns subjective taste into structured, computable signal that AI platforms, agents, and generators can act on.
Taste typicality
Taste typicality describes how characteristic an item is of a given taste, placing people and content on grounded perceptual scales instead of coarse categories.