ADEPT LION

---

Usually when we talk about pattern recognition, we mean the practice of identifying something consistent or repeatable within our rather complicated experience.  More recently, this pattern recognition may be in the context of machine learning, and a computer becomes the agent for identifying statistical patterns within very large datasets.  In either case, we are on a search for consistent elements that can produce a useful prediction regarding potential outcomes.  

ADEPT LION has a more assertive approach to pattern recognition.  Rather than waiting for  experience or data about the world to compose a model, it starts with a fixed method for modeling experience and data in the first place. It starts with a set of relationships, a set of rules governing these relationships, and a resultant set of patterns that can determine what, and how things might exist in the model in relationship to everything else.  These patterns mirror the meaning of what we observe and experience much like an analogy or a metaphor.  They are a means of expression - a pattern language - that always retain computability because the types of the relationships, rules and patterns never change.  They are always just patterns, entirely known by the method.  Thus, we can reason over models of many 'worlds' using the sense-making machinery that is baked-in to the capabilities of the pattern language.  This fixed way of defining pattern provides us with a consistently useful method for exploring complex matters of logic, language and information structure.