Artificial embarrassment

[X DdC]

Dear All & Alex Shkotin <alex.s...@gmail.com>

Yes:
    The current state of theoretical knowledge is that it's become
    very plentiful and quite tightly and intricately intertwined. On
    the other hand, almost all of it, in one form or another, exists
    in computers. This means it can be formalized.
But: How to formalize it and reason with it, that is what I was
referring to.

There is the detail of adding 'know how' to the regurgitation
ability of 'know what' -- as discussed by Terry Winograd in 1985 and
others for decades.
Notice the inabilities of the chat-systems to apply AI-algorithms and
dealing with the semantics of action-verbs in stories as lambasted in:
  Summary of 'Ontology families for AI-systems'
  https://www.ontooo.com/Ontology.pdf

    How should this vast array of knowledge be created, stored, and
    expanded?
I suggested to break down the problem by using the topic/domain
decompositions done by Steven Pinker and beyond.  Creating TBoxes for
these different topics can use the (120000?) concepts in Princeton's
Wordnet.

This is doable.  I created for a medical diagnosis system the TBox,
the ontology, the ABox with 658 diseases, 236 body locations, 102
body systems, 83 abnormal conditions, 928 symptoms, 153 external
causes.  And, of course, the diagnosis algorithm and additional
call-center functionality.  A patient simulator showed the ability to
identify each disease correctly.
See:
    "Medical Diagnosis Algorithm", in Conference WorldS4 2024 July
    on Intelligent Sustainable Systems, pages 121-130.
    https://mkscienceset.com/articles_file/455-_article1735797527.pdf
or: https://www.ontooo.pdf/Diagnosis.pdf
[ The diagnosis algorithm is generic and the design is language
agnostic.
This system should NOT be used by the general public; see the
paper why.]

     How to store it (mega ontology) & expand?
I suggested massive (domain) decompositions.

     My humble opinion would be a system of theory frameworks. How
     many will there be? I think about a hundred thousand.
No idea where that number is coming from.  Still, I agree that many
micro-ontologies are needed.

You appear to be in favor with your:
     One form of organizing projects of this class in an open format
     could be the nascent HG project Holon Graph | Community Groups |
     Discover W3C groups | W3C.
I am too far over the hill - with great views.  
Who can take over from here?

Knowing without doing is not OK.
Doing without knowing is dangerous.
AI needs to get its knowledge act together after 55 years.

dennis de champeaux

[John F Sowa]

Dennis,

As you might expect, I  strongly agree with your note to Ontolog Forum, and I recommend your article, which goes into more detail.   https://www.ontooo.com/Ontology.pdf .

I also agree with your comments about the Cyc system.  Lenat and I started in different directions, but we strongly agreed on most of the fundamental issues.  (1) A single universal ontology of everything is impossible.  (2) A simple general upper level ontology is useful, but it's the least important part of an ontology. (3) All the real work for any application depends on an open-ended hierarchy of smaller, special purpose modules, or microtheories, or applications, which may be reused or recombined or revised as needed,  (4).A systematic framework for supporting and relating these modules is essential.

For more about these issues, see Without Onrology LLMs are clueless, https://ontologforum.s3.amazonaws.com/OntologySummit2024/TrackA/LLMs-are-clueless--JohnSowa_20240228.pdf

There is also a YouTube of this talk, which received overe 12K downloads, but the slides are easier to read.  For the issues that Dennis discusses in his email and article, see slides 32 to the end. 

John

 

---

From: "X DdC" <ddc9...@gmail.com>

BGW-X%2BqKf4eaX_twA%40mail.gmail.com.

[Alex Shkotin]

Dear Dennis,

I think we are working in the same direction but with different emphasis. You take domains and create ontologies for them, while I take existing theories for a given domain and propose systematizing and formalizing their knowledge.

For a given area, the result will be roughly the same: systematized and concentrated knowledge.

For example, I'm absolutely in favor of applying Hoare's approach to algorithmic systems (your point 15).

To my taste (as I wrote in my previous letter), I read your titles like this: "14 Theories for research innovation?"

And TRIZ immediately comes to mind.

And so on for each of your chapters: for each of these topics, there are several theories, and they will have to be systematized and formalized, because "In IT, ontology is a formal theory that uses data."—see here.

This is a tremendous undertaking.

Mathematicians are also concerned with a similar problem—see

Let's do it.

My way is as follows: domain, theories, formalization (ontologies).

And AI is just a tool in this area of practice. But a deliverable is a formal theoretical knowledge in the form of a framework of units of knowledge, IMHO.

And about your Medical Ontology and Algorithm: is it available to the public? 

We don't create theories. We systematize and formalize existing ones. For each formula in the ontology there must be a reference to the expert source of this statement.

Alex

вт, 14 июл. 2026 г. в 01:11, X DdC <ddc9...@gmail.com>:

--
All contributions to this forum are covered by an open-source license.
For information about the wiki, the license, and how to subscribe or
unsubscribe to the forum, see http://ontologforum.org/info
---
You received this message because you are subscribed to the Google Groups "ontolog-forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontolog-foru...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ontolog-forum/CAJ%2B4qGdeuLjfTpF%3DbHkNjCkA6%3DVw3VnRkBGW-X%2BqKf4eaX_twA%40mail.gmail.com.