Things
A "thing" is an identifiable logical artifact such as an entity, idea, concept, event, situation, state, or process that is important to a system which you are concerned with. If you are not concerned with something, then that is "nothing" to your system of interest. But if it is important to you, then it is a "thing". Ideally, every thing can be unambiguously distinguished from every other thing in a system you are concerned with or any other thing in another system your system interacts with.
A system can be represented using a theory or model or conceptualization. Logic is a tool for describing a system. Logic systems are describable.
When you are constructing a theory or model, a thing is something that is represented within that theory or model. A thing is different than a token that is used to represent that thing in your theory/model. Ontology-like things can be used to physically describe things in a form readable to machines such as computers.
Things can be grouped or categorized or classified or typified into sets. These groupings or categorizations or classifications or typifying can be very useful and helpful when working with the things which make up a system.
Things can "be" parts of other things or things can "have" parts. Mereology is the theory of parthood relations: of the relations of part to whole and the relations of part to part within a whole. As such, assemblies of compound things can be made up of primitive things. Atomic design methodology describes these sorts of dynamics.
A taxonomy is a tool for organizing sets/groups/classes/categories of types. An ontology is another organization and description tool. Formally organizing something using a taxonomy or ontology provides unprecedented clarity.
Once things, types of things, assemblies of things, and constraints that assert or restrict how such assemblies can be constructed; you can create a knowledge representation that is machine-readable. Once you have the machine-readable knowledge representation, you can use a machine to reason using that representation.
But with such a system, there is always the threat of inaccuracy. Faults can occur and a component of the system might not act properly, per what is expected by system per the theory or model. These inaccuracies, faults, and other sorts of malfunctions can be problematic if they go undetected. But Lean Six Sigma principles, philosophies, and techniques can reduce or even eliminate inaccuracies, faults, and malfunctions.
Sure, the possibilities offered by artificial intelligence are more "sexy". But machine intelligence can only be useful if the threat/risk of inaccuracy, faults, malfunctions, and errors is properly managed.
Additional Information:
Comments
Post a Comment