Re: Knowledge representation for AI using domains or a universe of discourse

I'm not a member of IEEE to see the article.
However the Wikipedia page is very good.
One important point it makes is:
"Every ontology is a treaty–a social agreement among people with common motive in sharing." There are always many competing and differing views that make any general-purpose ontology impossible.

So what are we trying to achieve with knowledge representation, or even categorization of knowledge?
If we're truly trying to enable AI then maybe we could envisage a multi-step process (possibly with a dedicated AI agent for each, linked in some sort of on-trend agentic architecture):

  1.
Understanding of the problem to be solved in terms of its domain, granularity, and scope. This could even be decomposed into 3 sub-steps
    e.g. for problem of "how many miles between San Diego and Los Angeles", Domain = geospatial; Granularity =  city; scope = California
  2.
Find the appropriate representation(s) e.g. linked ontologies
  3.
Find the instance data (if not available for the representation it might require revistign step 2)
  4.
Transform the problem into a query on the data in terms of the representation
  5.
Transform the results into a form suitable for the user and for feeding back into local context memory

It seems to me that knowledge categorization could apply to steps 1 and 2.

HAGW
Pete

Pete Rivett (pete.rivett@federatedknowledge.com)
Federated Knowledge, LLC (LEI 98450013F6D4AFE18E67)
tel: +1-701-566-9534
Schedule a meeting at https://calendly.com/rivettp


________________________________
From: Milton Ponson <rwiciamsd@gmail.com>
Sent: Friday, June 27, 2025 4:19 PM
To: W3C AIKR CG <public-aikr@w3.org>; public-lod@w3.org <public-lod@w3.org>; public-cogai <public-cogai@w3.org>; paoladimaio10@googlemail.com <paoladimaio10@googlemail.com>
Subject: Fwd: Knowledge representation for AI using domains or a universe of discourse

I forgot the article link for the first abstract mentioned;
Granularity of knowledge, indiscernibility and rough sets
https://ieeexplore.ieee.org/document/687467


Milton Ponson
Rainbow Warriors Core Foundation
CIAMSD Institute-ICT4D Program
+2977459312
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean

---------- Forwarded message ---------
From: Paola Di Maio <paoladimaio10@gmail.com<mailto:paoladimaio10@gmail.com>>
Date: Fri, Jun 27, 2025, 09:23
Subject: Re: Knowledge representation for AI using domains or a universe of discourse
To: Milton Ponson <rwiciamsd@gmail.com<mailto:rwiciamsd@gmail.com>>
Cc: Owen Ambur <owen.ambur@verizon.net<mailto:owen.ambur@verizon.net>>, W3C AIKR CG <public-aikr@w3.org<mailto:public-aikr@w3.org>>, public-cogai <public-cogai@w3.org<mailto:public-cogai@w3.org>>, <public-lod@w3.org<mailto:public-lod@w3.org>>


Milton, Great to see you engage with the topic and with the categorization of knowledge

I am reminded that several knowledge categorization systems exist

It would be relevant to us to understand how the traditional knowledge categories that you cite
rooted in epistemoligy, relate to the knowledge representation for AI *which is the focus of this CG

The knowledge domain for AI KR is not the domain of 'knowledge categorization'
 it would be nonetheless be very relevant to KR, perhaps to explore and make explicit the relations between the two

PDM


On Thu, Jun 26, 2025 at 8:49 PM Milton Ponson <rwiciamsd@gmail.com<mailto:rwiciamsd@gmail.com>> wrote:
Excellent. And ChatGPT made recommendations about expanding, tweaking and refining domains in line with my review of my initial list. I think we can make additional refinements by adding a not exhaustive list of directives,  guidelines, legislation ( in casu EU, the Eu AI Act) currently existing in countries worldwide for ethical, open, accountable,  trustworthy etc AI use.

Milton Ponson
Rainbow Warriors Core Foundation
CIAMSD Institute-ICT4D Program
+2977459312
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean

On Wed, Jun 25, 2025, 23:10 Owen Ambur <owen.ambur@verizon.net<mailto:owen.ambur@verizon.net>> wrote:
Milton, while I can imagine there might be quite a lot of quibbling over the domains, this additional information is now included at https://stratml.us/docs/EARP.xml#uuid-1f8a0ae6-1b0a-4d1b-a201-5c8c9d1c8c01


ChatGPT suggests some potential modifications<https://chatgpt.com/share/685cb98a-dd68-800b-b751-971c639406f9>.

Owen Ambur
https://www.linkedin.com/in/owenambur/



On Wednesday, June 25, 2025 at 12:48:02 PM EDT, Milton Ponson <rwiciamsd@gmail.com<mailto:rwiciamsd@gmail.com>> wrote:


Thanks Owen,

I sense you were on to what I intend to do. Because the United Nations has the UNESCO, WIPO, ISO, and UNSD (UN Statistics Division) which together deal with various aspects of data and information, standardization and intellectual property rights issues, for scientific knowledge the process is more or less straightforward.

In the UN universe of discourse, knowledge is is a concept spanning many domains of discourse. Traditional, indigenous and orally transmitted knowledge is the first category.
It is currently being dealt with through SIL (https://www.sil.org),   UNESCO and several initiatives by organizations of indigenous peoples dealing with AI, traditional knowledge, data and knowledge sovereignty issues.

All the other categories of knowledge and data can be categorized either by universal subject coding sysyems, structured text components, or enumeration of informal text parameters. Semiotics, symbol sets, pictograms, petroglyphs and alphabets, spectral data are general categories which help define knowledge objects.

What needs to be done is to create a universal categorization for knowledge over the following domains of discourse:

  *   Mathematics;
  *   Computer Science;
  *   Physics and Astronomy;
  *   Chemistry, Biology and Biochemistry;
  *   Medical, Health and Life Sciences, including Genomics and Pharmacy;
  *   Earth Sciences; dealing with lithosphere, hydrosphere and atmosphere
  *   Marine Science including Oceanography;
  *   Engineering and Materials Sciences;
  *   Languages and Computational Linguistics;
  *   Philosophy;
  *   Social Sciences, including History, Political Sciences, Sociology and Anthropology;
  *   Art, Cultural and Creative Design Studies;
  *   Education, Vocational and Professional Training;
  *   Economics, Econometrics and Statistics;
  *   Political Sciences, and Law;
  *   Business and Public Sector Management.

Milton Ponson
Rainbow Warriors Core Foundation
CIAMSD Institute-ICT4D Program
+2977459312
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean

On Wed, Jun 25, 2025, 11:54 Owen Ambur <owen.ambur@verizon.net<mailto:owen.ambur@verizon.net>> wrote:
To help me make sense of this, I prompted ChatGPT to render a plan in StratML Part 2 format, which is now available at https://stratml.us/drybridge/index.htm#EARP or, more specifically, https://stratml.us/docs/EARP.xml


Owen Ambur
https://www.linkedin.com/in/owenambur/



On Tuesday, June 24, 2025 at 10:48:45 PM EDT, Paola Di Maio <paola.dimaio@gmail.com<mailto:paola.dimaio@gmail.com>> wrote:


Thank you Milton
Let us know how you propose to advance this suggestion
PDM

On Wed, Jun 25, 2025 at 6:27 AM Milton Ponson <rwiciamsd@gmail.com<mailto:rwiciamsd@gmail.com>> wrote:
Dear all,

We have reached the point in time where we must start to generalize the basis of knowledge representation for artificial intelligence.

In order to do so we must find (a) common denominator(s) for the different types of knowledge.

Assuming that the knowledge is digitized we can use the concept of the universe of discourse or more specifically domains of discourse.

These concepts lend themselves for use of set theory, type theory etc. and allow modeling and constructibility of consistent logical frameworks.

And they also allow the use of (hyper)graphs, algebraic geometry, category theory etc. to allow knowledge representation in ways that can be captured for data (structure) creation, data analysis and machine learning.

In addition we can use classification systems to uniquely identify domains of discourse as librarians do and as mathematicians do as well with the Mathematics Subject Classification.

This can lead to a much more efficient use of datasets, the use of which can be negotiated with the respective owners, avoiding the current intellectual property and data rights, data and knowledge sovereignty debates.

The indiscriminate use of scraping and unlimited production of internet content by artificial intelligence is contaminating available data.

After having examined the current use of artificial intelligence for solving sustainable development problems, both at global, regional and local levels, the use of (open) curated and/or standardized datasets is the only way to go.

This requires the use of a universe of discourse and more specifically domains of discourse.


Milton Ponson
Rainbow Warriors Core Foundation
CIAMSD Institute-ICT4D Program
+2977459312
PO Box 1154, Oranjestad
Aruba, Dutch Caribbean