walk-and-talk: DIKW pyramid/hierarchy

I walk in and around a park with my dog, talking about the the DIKW (Data, Information, Knowledge, Wisdom) class of models, eventually relating this to machine-centric science.

I've been thinking a bit more about the so-called DIKW conceptualization of, uh, going from data to information, to knowledge, to wisdom, whether it's a pipeline or a pyramid or whatever. Um, just a way of conceptually organizing, um, what we put in formation information.

Um, so I guess my, my latest thought on this is that, uh, data, uh, doesn't necessarily have to be structured. Um, what defines data is that it is used as input to a process or is output from a process. Um, so you can have things like unstructured data, meaning. Say free text. People are okay with that term unstructured data. They're generally not okay with the term unstructured information. And that's because information means in formation.

Um, it needs to be a form to it. It needs structure. Uh, so to go from, from data to information, uh, you, you know, you might have a data model. Which is to say, here are the various parts of this document or whatever, and this is a string and this is an in integer, et cetera. Um, but in order to go to information, you need an information model you need, well, these are the columns. These are what, what they mean. Um, you know, there are these, these fields and we'll organize it in a table or we'll organize it in a plot. This is what this axis means in these units. So the data is now information it's, it's structured.

Um, So I think, uh, once you do that enough times, and there are various, uh, methods by which you transform data to information like, like this is how I make this table. This is how I make this plot. This is how I. Um, collect these observations, um, and transform them, uh, either, you know, reshaping them or acre them or reducing them, uh, to some information that methodology now, um, can be discussed and presented on its own. And that's that's knowledge.

So the knowledge is, is. Methodological information, whereas, uh, information, uh, per se, uh, that's not methodological is, is situational information. It's observations. Um, Whereas methodological information is, uh, that's knowledge. Um, so you can say, well, these are the various methods by which I can, um, transform this data into this information.

Um, so that's that's knowledge. It takes a lot of iterations of going from data to information, um, to recognize that you have certain methods. Um, that you can document and you can document this knowhow, uh, this methodological information and that that's your knowledge. Um, finally going back, going up, um, you have wisdom.

So what is this about? I feel like wisdom is, um, you know, Rather than, uh, know how it's it's no, to it's, uh, knowing, uh, what to do. So you have a bunch of options for methodology and, um, you know, which, which one do you tend to apply in a given situation? Um, you know, you can, you can document those decisions and present those as information and, and that, uh, that would be the wisdom that would be the, the no, too.

So, uh, rather than your, your situational information, which is, I mean, your data information, uh, uh, To begin with, and you have your methodological information, which is your knowledge, this, this wisdom, this this know too, rather than the know how of knowledge, um, going from situational information to, uh, methodological information now it's, it's, it's philosophical information.

It's, um, epistemological. It's like, well, what do I actually accept as being knowledge, um, an appropriate how to in this situation, um, axiological, uh, information, uh, you know, what, what has value to be applied here? What methodology has value to be applied?

Um, also, uh, ontological, uh, information, you know, what, what do I want to be? What do I recognize as, as being the case and what do I want, um, to become, so it's, it's, it's. Um, knowing to, as opposed to knowing that and being able to document that. So you can see how all these build on each other, you kind of gather observational, um, , uh, data, um, you, you form it, um, according to some methodology that's that might be ad hoc and that's your information.

Eventually you can document and formalize this. These the suite of methodologies and that's your knowledge. Um, and as you do that enough times, eventually you can document and formalize the suite of, uh, decisions and, and the logic and, and, and conditional logic. Uh, that you use in order to apply the methodologies and that that's your wisdom that's um, so you're going from, from knowing that, to knowing how to knowing too.

Um, and I think that's one way of, of going up that DIKW data information, knowledge, wisdom, um, pyramid hierarchy, pipeline, what you will, uh, and, uh, Yeah, you can iterate at any point of this. And if you do so in a, in an overall scientific matter, uh, scientific manner, rather where you, um, you make deductions, you make inductions, you make abductions, so you kind of explore the hysteresis from deduction to induction and make hypotheses and, and test them.

Uh, so the overall method. On faith is, is, is, is the scientific method. Um, this is how you can, uh, be machine centric about, about your science, um, machine centric, because you're leaning on the machine to execute and keep you honest on these formalisms that you make about. Well, this is my information. This is my knowledge. This is my wisdom. Um, you can, you can encode those so that, uh, computer can understand them.

Um, as Abelson and Sussman have said, it's, it's, it's really for humans to read it only incidentally for machines to execute, but, uh, the fact that you're making it able, you're making it able, you're making a machine able to execute it. Uh, that's, that's really important to keep you, keep you honest. Um, in coming up with and accreting this, um, This knowledge and this wisdom, this, this, this repository of know-how and this repository of know-to. Um, and of course, a repository of, of repo, various repositories of, of, of know-that, um, uh, using information models,

um, the data isn't, uh, important in that it stands on its own. The data again is, is sort of input to these processes. The, the things that you actually serve up or. These fact databases with information models or knowledge models or, um, wisdom models, I guess, philosophical models, ontological, axiological, and epistemological models. Um, the data is, is, is the raw stuff that, um, It's important to point out, you know, for, for provenance reasons. But, um, you know, data is very, um, dependent on the context. Okay. I think I've babbled on enough for now.