If traditional philosophy/ontology is dead, we need data ontology

[Azamat Abdoullaev]

Speaking to Google’s Zeitgeist Conference in Hertfordshire, Hawking said that fundamental questions about the nature of the universe could not be resolved without hard data such as that currently being derived from the Large Hadron Collider and space research. “Most of us don’t worry about these questions most of the time. But almost all of us must sometimes wonder: Why are we here? Where do we come from? Traditionally, these are questions for philosophy, but philosophy is dead”.

“Philosophers have not kept up with modern developments in science. Particularly physics.”

My answers:

https://www.quora.com/How-does-model-dependent-realism-prove-that-philosophy-is-dead-Why-should-one-accept-the-claim-that-philosophy-is-dead/answer/Kiryl-Persianov?comment_id=237009271&comment_type=2&__filter__=all&__nsrc__=notif_page&__sncid__=21910010862&__snid3__=30134738096

https://www.linkedin.com/pulse/god-dead-philosophy-science-rise-real-ai-azamat-abdoullaev/?trackingId=QssyH2lxQn2ScrtYCxQL%2Bg%3D%3D

[Michael DeBellis]

I kind of agree with Hawking but not completely.  IMO, there are people who write about philosophical issues but most of them aren't in philosophy departments but in Anthropology, Psychology, Linguistics, Physics, etc.. Patricia Churchland wrote a fascinating book called Neurophilosophy a few decades ago. There is an anthropologist at UC Berkeley named Terrance Deacon whose work is very much influenced by and has repercussions for philosophy. Dawkins' God Delusion book had many references to and implications for philosophy (the issue of God of course but he also talked about other philosophical issues and he still does, such as his article on why essentialism is a fundamental error) And of course Chomsky. 

I agree with Hawking that what goes on in most American philosophy departments is mostly pseudoscience and not even wrong. The problem is that there is an artificial barrier that most modern American academic philosophers believe in between science and philosophy. Virtually none of the people we think of as great philosophers like Hume, Kant, Descartes, Leibniz, Frege, even Nietzsche recognized such a distinction and many of them did significant work in areas that are now considered Math and Science as well as their philosophical work. And for them (as Chomsky points out) there was no significant distinction between their scientific work and their philosophical work. Good philosophy in the tradition of people like Hume and Socrates isn't dead, it's just mostly not done in academic philosophy departments. 

Michael

[Dr. Lars Ludwig]

Dear Azamat, 

I think the scientific problems you describe reflect real core problems of current science.

Azamat> "... total convergence of knowledge, technology and people, Trans-AI = Narrow AI, ML, DL + Symbolic AI + Human Intelligence)."

What you paint (with broad strokes) as total convergence of "knowledge, technology, and people" in "trans-AI" is actually very similar to my ideas about "extended artificial memory". It's a form of re-aligning technology with memory and ways of extending mind in a more controlled way (which would comprise "narrow AI, ML, DL, symbolic AI, human intelligence").  (https://d-nb.info/1045194794/34) 

I think we literally need a Gutenberg moment, a leap in technology to kickstart transdisciplinary . Language as a medium was extended by the Gutenberg press in quality (exact replication), quantity (countless replications), and time (enduring replications).

I have a possible crystallization point in mind: We could try to tackle language as an unprecise, ambiguous medium. This time, the extension needs to aim at 'precision of expression' and thus comprehesibility. Traditional natural language is ambiguous; and mathematical and logic formalization are insufficient surrogat media of universal precise expression and communication. Ontological technologies and representation media are 'stiff' (their UIs tend to be horrible, if at all usable), and (even though John Sowa will not agree) existential graphs and similar semantic graphic-oriented representations are difficult and largely non-intuitive to read (dead ends).

Transparent extended unambiguous or less ambiguous language technology ought to be augmented and must be thought-accompanying and thought-extending in expression, and automatically (sub-consciously) readable. - I think (after some self-experimentation) the best path would be a medium using an exendable basic set of semantically and logically disambiguated (serially) nested language chunks. Those are formal representational constructs most similar to natural langage (memory) structures; they could help enhance (ontologize) communication and would thus also increasingly consolidate the individual and societal 'mnemotom'.   

Best, 

Lars               

 

    

[Adrian Walker]

Dr. Ludwig.

You wrote :

I think we literally need a Gutenberg moment, a leap in technology to kickstart transdisciplinary . Language as a medium was extended by the Gutenberg press in quality (exact replication), quantity (countless replications), and time (enduring replications).

I have a possible crystallization point in mind: We could try to tackle language as an unprecise, ambiguous medium. This time, the extension needs to aim at 'precision of expression' and thus comprehesibility. Traditional natural language is ambiguous semico and mathematical and logic formalization are insufficient surrogat media of universal precise expression and communication. Ontological technologies and representation media are 'stiff' (their UIs tend to be horrible, if at all usable), and (even though John Sowa will not agree) existential graphs and similar semantic graphic-oriented representations are difficult and largely non-intuitive to read (dead ends).

 It's possible that what we call Executable English (EE) fits the requirements for "extension needs to aim at 'precision of expression' and thus comprehensibility"

The  EE notation consists of syllogism-like rules such as listed in https://www.executable-english.com/rule_examples.html .  The rules look similar to classical expert systems, but they are interpreted differently -- particularly wrt to negation.

A system that demonstrates how this works for precision of expression is live on the web at https://www.executable-english.com with many examples that one can view, run and change using a browser.  One can also write  and run new examples, using words and phrases that have meaning for non-technical users, without any need for vocabulary or grammar maintenance.  

Of course there is a tradeoff involved,  but the main point is that the approach could perhaps contribute to precision of expression for data-ontology/general AI designs of the future.

Just some thoughts.   -- Adrian

Adrian Walker
Executable English LLC  
San Jose, CA, USA
USA 860-830-2085 (California time)
https://www.executable-english.com

[Dr. Lars Ludwig]

Dear Adrian, 

Adrian> It's possible that what we call Executable English (EE) fits the requirements for "extension needs to aim at 'precision of expression' and thus comprehensibility"

I think any natural-language-oriented ontology (technology, system) providing more than aliases or some natural language labels for a concept and being translatable into a logical notation points into the direction of "more precision", and many existing systems will contain ideas that could be re-used in the type of system I envision. - By the way, "executable English" is a nice name.

The examples I see on the web page you linked remind me of controlled natural language (systems). What I have in mind (I know I am guilty of so far not having delivered substantial material for clarification) is much closer to actual natural language (production) and would be a very advanced (highly interactive) editor/reader UI.

I have myself developed a (merely) long-time-experimental system and an associated ontology (of now nearly 1 million natural language chunks / linear units / concepts) and starting to get an intuition of how the reading of largely pre-chunked texts considerably changes (enhances) the experience of reading; and I start to be able to quickly express/formalize ever more complex arbitrary thought based on more primitive chunks. 

Best, 

Lars

[Alex Shkotin]

Colleagues,

Great topic! Let me add my two pence:

-Eglish is a formal language (R. Montague) we just lost a way how to get a structure behind the text. For example, let's take "A man works." from another thread and write it formally ((a man) (walk s)) where "a" and "s" are unary operators - modifiers, and modified unary operator (walk s) is applied postfixly to a modified entity (a man) . We get a structure at least:-)

-terms of natural language are extremely polymorphic i.e. have many meanings and even ways of usage: in English word may be easily a verb and noun. But any meaning has a definition and a "context" of full definition is only a theory where it has been issued. Where theory is a set of axioms and derivation rules. The theory is useful if its models match reality.

Best,

Alex

пт, 17 дек. 2021 г. в 01:57, Dr. Lars Ludwig <ma...@lars-ludwig.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 on the web visit https://groups.google.com/d/msgid/ontolog-forum/821249863.390760.1639695448545%40email.ionos.de.

[Dr. Lars Ludwig]

Hi Alex, 

I like your example ("a man walks").

I typed "a man walks" into my system and did 1 click (doing chunking):

And this (above) is the linear unit (concept of a verbal phrase) I ended up with. Without going into detail, it basically states that a man - a noun word phrase including an singular adnoun denoting a singular male person that is a living thing is related to an activity of walking - 'walks' being a singular verb form of an activity / doing. 

To clarify the relation between 'a man' and the verb I specified this with another click: 

 

As this is only one phrase of a potential natural language concept syn-set (ignoring some syntacto-semantic variations), I added another click to come up with: 

So after 3 clicks I ended up with a fully differentiated natural language / linear units syn-set of a verbal phrase construct including 13 entities.  

This gives you an idea of what is possible. Of course, all of this could be automated utilizing ML models over existing structures in a large linear units set (which would very likely have provided the same result, just leaving some uncertainty).

Best, 

Lars 

[Alex Shkotin]

Hi Lars,

When JFS pointed out to his great CL I am asking just how to get and install compiler, interpreter, etc. I know how to do this for Coq, for example. It is great to have a standard, but we can't put it on the bread:-)

If you have Information System the question is if it is available for the public, as EE of Adrian Walker, for example.

I do not understand your notation so I need documentation too:-)

In the area of IT there are usual anticipations, you know.

What I am sure from your letter - your system does work - it's great!

Alex

пт, 17 дек. 2021 г. в 12:32, Dr. Lars Ludwig <ma...@lars-ludwig.com>:

[Michael DeBellis]

Alex, I disagree that English is a formal language. A formal language (propositional logic, FOL, modal logic) must have a well defined syntax that can be parsed without ambiguity. If I say: "I saw the man on the hill with a telescope" you can parse that multiple ways. E.g., meaning "I used a telescope to see the man on the hill" or "I saw the man on the hill and he had a telescope". I also don't agree that Montague claimed English was a formal language. He created a formal language to represent the meaning of English sentences (similar to what Chomsky calls the Logical Form in his theory) but never claimed English itself was a formal language, at least from what I remember. 

Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/ssb8bczkoig/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROSgoUsx%2BN2zzE5GumXnnqAYggRDEzUv%3D-OHdb9Qf8dhsQ%40mail.gmail.com.

[Alex Shkotin]

Dear Barbara,

Could you please clarify for us if R. Montague claimed that English is a formal language?

Sincerely yours,

Alex Shkotin

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFike9bUKQu%3DvRXX6%2BCvEs%3DBbi1mJOTrQv%3DzQT1suRyz5JyQ%40mail.gmail.com.

[Michael DeBellis]

On another thread I used one of the classic Chomsky examples of ambiguous syntax: "I saw the man on the hill the a telescope" I would be interested to know how the system described in previous posts in this thread handles that sentence. Parsing short non-ambiguous sentences like "a man walks" is fairly trivial. I've written parsers for such sentences a couple times, once using a tool called Refine to define grammars and parsers and once with just the rules in the Knowledge Engineering Environment (KEE).  I could parse very simple sentences but as soon as I had sentences with anaphora and other common problems it became orders of magnitude more complex. 

[Alex Shkotin]

Michael,

It is clear that if English is a formal language it is a kind of HOL. Details are the research area.

Ambiguity in formal languages is solved the same way is in natural: show variants and ask to choose.

The way of Chomsky is not the best - see the better way here [1].

Alex

[1] https://www.researchgate.net/publication/45865036_Program_structure

пт, 17 дек. 2021 г. в 18:29, Michael DeBellis <mdebe...@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFike9bUKQu%3DvRXX6%2BCvEs%3DBbi1mJOTrQv%3DzQT1suRyz5JyQ%40mail.gmail.com.

[Alex Shkotin]

good idea. Let's add ACE to the test.

And my favorite tools are Bison and flex.

пт, 17 дек. 2021 г. в 19:33, Michael DeBellis <mdebe...@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 on the web visit https://groups.google.com/d/msgid/ontolog-forum/fcce8a17-c756-4041-9933-d3ffe8acec1an%40googlegroups.com.

[Michael DeBellis]

Alex, I think I was wrong about Montague. I apologize.  I found a paper of his called English as a Formal Language. But regardless of what Montague thought I have read a fair amount of modern Linguistics and I have never heard anyone claim that English or other natural languages are formal languages. 

In fact most Linguists when they talk about it are adamant to point out that this is absolutely not the case. In one of his books (I think Women, Fire, and Dangerous Things) Lakoff spends a great deal of time arguing that natural languages are not the same as formal languages which he says is one of the major problems with Chomsky's approach. Which really made my head spin because if you read just about anything that Chomsky has written he is just as adamant that natural languages are not formal languages. 

There was a time when Anglo-American philosophers attempted to use the same model theoretic semantics as is used for FOL on the semantics of natural languages. I think that is where Montague was coming from. But to my knowledge there are no major Linguists who still think that natural languages as formal languages is coherent for the reason I mentioned before. When you learn any formal language you learn rules of syntax. What is a Well Formed Formula. And a WFF is always unambiguous to parse. There is just one possible syntax. That is what makes natural languages so complex is that not only is the semantics context dependent (as it is with formal languages) but even the syntax is context dependent (as it is not with formal languages). 

Michael

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxROT0ytmi5WeczNZYRwauRt2Zm-Wup8B9s_3sEUduc9pKTw%40mail.gmail.com.

[Alex Shkotin]

As Barbara Partee is not a member of ontolog-forum let me forward her response.

---------- Forwarded message ---------
От: Barbara Partee <par...@linguist.umass.edu>
Date: пт, 17 дек. 2021 г. в 19:51
Subject: Re: [ontolog-forum] Re: If traditional philosophy/ontology is dead, we need data ontology

To: Alex Shkotin <alex.s...@gmail.com>
Cc: ontolog-forum <ontolo...@googlegroups.com>

Dear Alex,

    Thanks for inviting me into this discussion. I’ll be happy to compose a reply today or tomorrow. The short answer is yes he did, more exactly as represented in the first sentence of his "English as a Formal Language", “I reject the contention that an important difference exists between formal and natural languages."

   But I'll expand on that slightly to say a few words about 'surface ambiguity', which seems to be the main reason that people think that English (and other natural languages) couldn't be a formal language.

    So, more soon and thanks again for asking.

Best

Barbara

[Alex Shkotin]

Michael,

The sentence in the mind is not a linear structure. We need linearization to speak it loudly (J. Corcoran) to another person. And sometimes there is more than one mind-structure having the same linearisation. 

Is the sentence in the mind is formal i.e. its structure is under strict rules? I hope so. 

Alex

пт, 17 дек. 2021 г. в 20:00, Michael DeBellis <mdebe...@gmail.com>:

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CALGFikdGxiYkCyDZQsDnkCu10pTDQMPizOehD9s3b5epxPW2PQ%40mail.gmail.com.

[Dr. Lars Ludwig]

Dear Michael, 

> On another thread I used one of the classic Chomsky examples of ambiguous syntax: "I saw the man on the hill the a telescope" I would be interested to know how the system described in previous posts in this thread handles that sentence. Parsing short non-ambiguous sentences like "a man walks" is fairly trivial.

It took me a few minutes and the result could be improved visually, but formally these 5 linear units are totally disambiguated (I guess one could construct even more):

I saw || (the man [being] on the hill) (with [him having] a telescope)

I saw || (the man [being]) (on the hill with [it having] a telescope)

 

(I saw (the man [where was he] on the hill)) || with the aid of a telescope.

(I saw the man) [where was I when I saw him?] (on the hill) || with the aid of a telescope

I saw the man || [where did this take place] ]on the hill with [having] a telescope

Best, 

Lars 

Alex Shkotin <alex.s...@gmail.com> hat am 17.12.2021 17:42 geschrieben:

To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/CAFxxRORhFsg73O3Jm_obOsS-QwqozhCuO%3D3BgVOwusc_o%3D%2BDmA%40mail.gmail.com.

[Dr. Lars Ludwig]

Sorry, I had a copy-paste error: the last one is: 

Dr. Lars Ludwig <ma...@lars-ludwig.com> hat am 17.12.2021 22:27 geschrieben:

[John F Sowa]

Alex, Michael, and Matthew,

 

I'm writing a longer note about  the fundamental issues we have been discussing.  But as just a quick  overview of the issues about the limitations of formal logics to represent full natural language semantics, see the first 20 slides of http://jfsowa.com/talks/natlog.pdf .  You can read more if you wish, but please look at those 20 slides.   There are many pretty pictures in those slides, nearly all of which were drawn by other people (whom I cite).

 

Re mathematics:  All possible patterns of anything and everything can be represented by some version of mathematics.   --  but you may need to use an open-ended variety of math, including versions that have not yet been discovered.  For evidence that math is sufficiently general to represent everything we can see or hear, note that all TV and movie cameras record  linear streams of bits that can be mathematically transformed to and from 4-D space-time presentations with stereophonic sound.

 

For a truly universal ontology,, you need to represent all possible mathematical theories, since it is impossible to know in advance what kinds of strange things may exist in our universe (or perhaps multiverse).

 

Just consider, for example, my earlier note about living robots that can reproduce.  That discovery challenges a huge number of hypotheses  and ontologies about living things and robots.

    

Summary:  We need either an open-ended number of ontologies or an open-ended number of branches under the top node called Entity.  Therefore, we need some powerful methods for relating different ontologies or different branches of any single ontology.  Therefore, we cannot expect humans to specify anything more than small, local aspects of any general purpose ontology.  Therefore, we need powerful new methods for developing, maintaining , and relating ontologies.

 

Conclusion:  As I have said for years, the ISO standard for TLOs is not just bad, it is hopelessly bad.  

 

And by the way, the word 'data' in the subject line above is also hopelessly misleading.  All computer data is mathematical -- just look inside any computer.   Solution:  Replace "data ontology" with "mathematics".

 

John

[doug foxvog]

The images came out garbled on my computer.

S = "I"
V = "saw"
O' = "the man"
PP1 = "on the hill"
PP2 = "with a telescope"

But, the sentence is SVO with two prepositional phrase, i.e., S V O' PP1
PP2. The question is what PP1 and PP2 modify. Is O equal to O', (O' +
PP1), (O' + PP1 + PP2), or (O' + (PP1+PP2)). English wouldn't skip over a
prepositional phrase in modification, so O != (O' + PP2) with PP1
modifying S.

Standard English has prepositional phrases <i>normally</i> modify the
first preceding phrase that it would appropriately modify.

Can PP1 modify O'? Yes, so it should.

Can PP2 modify the object of PP1? It's possible, but questionable. I
note that the reference is to a particular hill. PP2 might help determine
which particular hill. For example if the hill were Mount Palomar, then
PP0 = PP1 + PP2 = "on the (hill with a telescope)"
O = O' + PP0 = "the (man on the (hill with a telescope))"

> I saw || (the man [being]) (on the hill with [it having] a telescope)

alt: I saw || (the man [being] (on the hill with [it having] a telescope))

English is normally stated in ways that are not ambiguous. If PP1 and PP2
were both part of O (i.e, they both modify O') then phrasing the sentence
S V O' PP2 PP1 would clearly have
O = O' + PP2 +PP1 = (a (man (with a telescope) (on the hill)))

> I saw || (the man [being] on the hill) (with [him having] telescope)

The non-use of this form suggests that P2 may not modify O'. This
suggests that P2 may modify the component before O', V. Thus the Verb
Predicate would be:
VP = (V + PP2 (O' + PP1)) = (saw (with a telescope) (a man on the hill)"
This would be more clearly stated as:

PP2 + S + V + (O' + PP1)

> (I saw (the man [where was he] on the hill)) || with the aid of a
telescope.

Can PP1 modify S? Not unless PP2 modifies S or PP1. But both of these
would be awkward.

This would be more likely stated as
PP1 PP2 S V O'

> (I saw the man) [where was I when I saw him?] (on the hill) || with
> the aid of a telescope

or
PP1 S PP2 V O'

> I saw the man || [where did this take place] ]on the hill with
> [having] a telescope

-- doug

On Fri, December 17, 2021 16:29, Dr. Lars Ludwig wrote:
> <!doctype html>
> <html>
> <head>
> <meta charset="UTF-8">
> </head>
> <body>
> <div class="default-style">
> Sorry, I had a copy-paste error: the last one is:&nbsp;
> </div>
> <div>
> <img src="cid:06bdd3379b274375...@open-xchange.com"
> style="max-width: 100%;" class="aspect-ratio" alt=""
> from_clipboard="true">
> <br>
> </div>
> <blockquote type="cite">
> <div>
> Dr. Lars Ludwig &lt;ma...@lars-ludwig.com&gt; hat am 17.12.2021 22:27
> geschrieben:
> </div>
> <div>
> <br>
> </div>
> <div>
> <br>
> </div>
> <div class="default-style">
> Dear Michael,&nbsp;
> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> &gt; On another thread I used one of the classic Chomsky examples of

> ambiguous syntax: "I saw the man on the hill the a telescope" I would
> be interested to know how the system described in previous posts in
> this thread handles that sentence. Parsing short non-ambiguous
> sentences like "a man walks" is fairly trivial.

> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">

> It took me a few minutes and the result could be improved visually,
> but formally these 5 linear units are totally disambiguated (I guess
> one could construct even more):

> </div>
> <div class="default-style">
> <br>
> </div>
> <div>
> <img alt="" class="aspect-ratio" style="max-width: 100%;"
> src="cid:90bd4b65c5d04a3d...@open-xchange.com">
> <br>
> </div>
> <div class="default-style">

> I saw || (the man [being] on the hill) (with [him having] a
> telescope)

> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> <img alt="" class="aspect-ratio" style="max-width: 100%;"
> src="cid:d1f4e77c907f41a8...@open-xchange.com">
> <br>
> </div>
> <div class="default-style">
> <div class="default-style">

> I saw || (the man [being]) (on the hill with [it having] a
> telescope)

> </div>
> </div>
> <div class="default-style">
> &nbsp;
> <img alt="" class="aspect-ratio" style="max-width: 100%;"
> src="cid:2710c157bb5f46a7...@open-xchange.com">
> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">

> (I saw (the man [where was he] on the hill)) || with the aid of a
> telescope.

> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> <img alt="" class="aspect-ratio" style="max-width: 100%;"
> src="cid:4060928bc3404e82...@open-xchange.com">
> <br>
> </div>
> <div class="default-style">

> (I saw the man) [where was I when I saw him?] (on the hill) || with
> the aid of a telescope

> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> <img alt="" class="aspect-ratio" style="max-width: 100%;"
> src="cid:e893de5c0abc4e7f...@open-xchange.com">
> <br>
> </div>
> <div class="default-style">

> I saw the man || [where did this take place] ]on the hill with
> [having] a telescope

> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> Best,&nbsp;
> </div>
> <div class="default-style">
> <br>
> </div>
> <div class="default-style">
> Lars&nbsp;
> </div>
> <blockquote type="cite">
> <div>
> Alex Shkotin &lt;alex.s...@gmail.com&gt; hat am 17.12.2021 17:42
> geschrieben:
> </div>
> <div>
> <br>
> </div>
> <div>
> <br>
> </div>
> <div dir="ltr">
> <div class="gmail_default" style="font-family: georgia,serif;">
> good idea. Let's add <a
> href="https://en.wikipedia.org/wiki/Attempto_Controlled_English">ACE</a>
> to the test.
> </div>
> <div class="gmail_default" style="font-family: georgia,serif;">
> And my favorite tools are <a
> href="https://en.wikipedia.org/wiki/GNU_Bison">Bison</a> and flex.
> </div>
> </div>
> <br>
> <div class="gmail_quote">
> <div dir="ltr" class="gmail_attr">
> пт, 17 дек. 2021 г. в 19:33, Michael DeBellis &lt;<a
> href="mailto:mdebe...@gmail.com">mdebe...@gmail.com</a>&gt;:
> <br>
> </div>
> <blockquote>

> On another thread I used one of the classic Chomsky examples of
> ambiguous syntax: "I saw the man on the hill the a telescope" I
> would be interested to know how the system described in previous
> posts in this thread handles that sentence. Parsing short
> non-ambiguous sentences like "a man walks" is fairly trivial. I've
> written parsers for such sentences a couple times, once using a
> tool called Refine to define grammars and parsers and once with
> just the rules in the Knowledge Engineering Environment

> (KEE).&nbsp; I could parse very simple sentences but as soon as I

> had sentences with anaphora and other common problems it became

> orders of magnitude more complex.&nbsp;
> <br>
> <br>
> <div class="gmail_quote">
> <div dir="auto" class="gmail_attr">

> On Friday, December 17, 2021 at 1:32:35 AM UTC-8 mail wrote:

> <br>
> </div>
> <blockquote> <u></u>
> <div>
> <div>
> Hi Alex,&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>

> I like your example ("a man walks").

> </div>
> <div>
> <br>
> </div>
> <div>

> I typed "a man walks" into my system and did 1 click (doing
> chunking):

> </div>
> <div>
> <br>
> </div>
> <div>
> <img
> src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.1&amp;view=1"
> style="max-width: 100%;" alt="">
> </div>
> <div>

> And this (above) is the linear unit (concept of a verbal
> phrase) I ended up with. Without going into detail, it
> basically states that a man - a noun word phrase including an
> singular adnoun denoting a singular male person that is a
> living thing is related to an activity of walking - 'walks'

> being a singular verb form of an activity / doing.&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>

> To clarify the relation between 'a man' and the verb I

> specified this with another click:&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>
> <img
> src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.2&amp;view=1"
> style="max-width: 100%;" alt="">&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>

> As this is only one phrase of a potential natural language
> concept syn-set (ignoring some syntacto-semantic variations), I

> added another click to come up with:&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>
> <img
> src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.3&amp;view=1"
> style="max-width: 100%;" alt="">
> <br>
> </div>
> <div>

> So after 3 clicks I ended up with a fully differentiated
> natural language / linear units syn-set of a verbal phrase

> construct including 13 entities.&nbsp;&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>

> This gives you an idea of what is possible. Of course, all of
> this could be automated utilizing ML models over existing
> structures in a large linear units set (which would very likely
> have provided the same result, just leaving some uncertainty).

> </div>
> <div>
> <br>
> </div>
> <div>
> Best,&nbsp;
> </div>
> <div>
> <br>
> </div>
> <div>
> Lars&nbsp;
> </div>
> <blockquote type="cite">
> <br>
> </blockquote>
> </div>
> </blockquote>
> </div>

[John F Sowa]

Dear Lars,

 

Please do not make any assumptions about what I think or believe.    On the following point, I agree with you:

 

Lars>  and (even though John Sowa will not agree) existential graphs and similar semantic graphic-oriented representations are difficult and largely non-intuitive to read.

 

Different people have different needs, goals, and expertise.  As the subject line says, "Different strokes for different folks."  Existential graphs are a very general and powerful theoretical foundation.  But I would never recommend them for a user interface.  I wouldn't even use them myself for that purpose.

 

But I disagree with the following point for the same reason (Different strokes for different folks):

 

Lars> I think (after some self-experimentation) the best path would be a medium using an exendable basic set of semantically and logically disambiguated (serially) nested language chunks..

 

I have collaborated with some colleagues in developing some advanced AI applications.  Different users require totally different interfaces, and we accept their specifications (whatever they may be) and support whatever requirements they ask for.

 

See the applications in the following slides:  http://jfsowa.com/talks/cogmem.pdf  (Note that Arun Majumdar is the first aurhor.)

 

Prolog is the primary language for implementing the system.  Conceptual graphs (an extension of existential graphs for representing natural languages) is the primary knowledge representation language. And Cognitive Memory (CM) is the key tool for supporting advanced AI technology.  Arun originally implemented CM in Prolog, but he rewrote important parts of it in C for better performance.

 

There is much more to say, but I recommend those slides for examples of what can be done.  Those applications were implemented over 10 years ago.  We have more advanced tools now, but even the projects in cogmem.pdf are far more advanced than anything that can be done with the Semantic Web tools. 

 

John

[Dr. Lars Ludwig]

Dear Doug, 

The texts added to the images (of the HTML code generated visualizing the formalizations in the ontology)  were quickly made descriptions (and I am guilty of not having put much effort into them). The descriptive ASCII-formalizations you use (e.g. "PP2 + S + V + (O' + PP1)") are clearer (structured), just not maintaining the sequence of phrases stated in the original sentence (which had been my dubious intention in the descriptions). 

As to whether the sequence of prepositional phrases is common or well-formed (w.r.t. the different sentence interpretations I formalize) in standard english: I don't really care, as whatever way you could read (understand) the sentence is fine with me and the/a system ought to be able to formalize and disambiguate it in that sense using the surface expression given (I don't blame a simple man or master yoda for their languages ;-) 

Best, 

Lars   

doug foxvog <do...@foxvog.org> hat am 18.12.2021 05:51 geschrieben:

</div>

<div>

<img src="cid:06bdd3379b274375...@open-xchange.com"

style="max-width: 100%;" class="aspect-ratio" alt=""

from_clipboard="true">

<br>

</div>

<blockquote type="cite">

<div>

Dr. Lars Ludwig <ma...@lars-ludwig.com> hat am 17.12.2021 22:27

geschrieben:

</div>

<div>

<br>

</div>

<div>

<br>

</div>

<div class="default-style">

Dear Michael, 

</div>

<div class="default-style">

<br>

</div>

<div class="default-style">

</div>

<div class="default-style">

<br>

</div>

<div class="default-style">

Lars 

</div>

<blockquote type="cite">

<div>

Alex Shkotin <alex.s...@gmail.com> hat am 17.12.2021 17:42

geschrieben:

</div>

<div>

<br>

</div>

<div>

<br>

</div>

<div dir="ltr">

<div class="gmail_default" style="font-family: georgia,serif;">

good idea. Let's add <a

href="ACE" rel="noopener" target="_blank">https://en.wikipedia.org/wiki/Attempto_Controlled_English">ACE</a>

to the test.

</div>

<div class="gmail_default" style="font-family: georgia,serif;">

And my favorite tools are <a

href="Bison" rel="noopener" target="_blank">https://en.wikipedia.org/wiki/GNU_Bison">Bison</a> and flex.

</div>

</div>

<br>

<div class="gmail_quote">

<div dir="ltr" class="gmail_attr">

пт, 17 дек. 2021 г. в 19:33, Michael DeBellis <<a

href="mailto:mdebe...@gmail.com"mdebe...@gmail.com" data-mce-href="mailto:>mdebe...@gmail.com">>mdebe...@gmail.com</a>>:

<br>

</div>

<blockquote>

On another thread I used one of the classic Chomsky examples of

ambiguous syntax: "I saw the man on the hill the a telescope" I

would be interested to know how the system described in previous

posts in this thread handles that sentence. Parsing short

non-ambiguous sentences like "a man walks" is fairly trivial. I've

written parsers for such sentences a couple times, once using a

tool called Refine to define grammars and parsers and once with

just the rules in the Knowledge Engineering Environment

(KEE).  I could parse very simple sentences but as soon as I

had sentences with anaphora and other common problems it became

orders of magnitude more complex. 

<br>

<br>

<div class="gmail_quote">

<div dir="auto" class="gmail_attr">

On Friday, December 17, 2021 at 1:32:35 AM UTC-8 mail wrote:

<br>

</div>

<blockquote> <u></u>

<div>

<div>

Hi Alex, 

</div>

<div>

<br>

</div>

<div>

I like your example ("a man walks").

</div>

<div>

<br>

</div>

<div>

I typed "a man walks" into my system and did 1 click (doing

chunking):

</div>

<div>

<br>

</div>

<div>

<img

src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.1&view=1"

style="max-width: 100%;" alt="">

</div>

<div>

And this (above) is the linear unit (concept of a verbal

phrase) I ended up with. Without going into detail, it

basically states that a man - a noun word phrase including an

singular adnoun denoting a singular male person that is a

living thing is related to an activity of walking - 'walks'

being a singular verb form of an activity / doing. 

</div>

<div>

<br>

</div>

<div>

To clarify the relation between 'a man' and the verb I

specified this with another click: 

</div>

<div>

<br>

</div>

<div>

<img

src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.2&view=1"

style="max-width: 100%;" alt=""> 

</div>

<div>

<br>

</div>

<div>

As this is only one phrase of a potential natural language

concept syn-set (ignoring some syntacto-semantic variations), I

added another click to come up with: 

</div>

<div>

<br>

</div>

<div>

<img

src="https://groups.google.com/group/ontolog-forum/attach/b83714c177e59/image.png?part=0.3&view=1"

style="max-width: 100%;" alt="">

<br>

</div>

<div>

So after 3 clicks I ended up with a fully differentiated

natural language / linear units syn-set of a verbal phrase

construct including 13 entities.  

</div>

<div>

<br>

</div>

<div>

This gives you an idea of what is possible. Of course, all of

this could be automated utilizing ML models over existing

structures in a large linear units set (which would very likely

have provided the same result, just leaving some uncertainty).

</div>

<div>

<br>

</div>

<div>

Best, 

</div>

<div>

<br>

</div>

<div>

Lars 

</div>

<blockquote type="cite">

<br>

</blockquote>

</div>

</blockquote>

</div>

--

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 on the web visit https://groups.google.com/d/msgid/ontolog-forum/aa54af16259e873a4e765039e5df4ba5.squirrel%40emailmg.ipage.com.

[Dr. Lars Ludwig]

Dear John, 

> Please do not make any assumptions about what I think or believe.    

You are right in that I was wrong in this particular assumption.  

> But I disagree with the following point for the same reason (Different strokes for different folks): Lars> I think (after some self-experimentation) the best path would be a medium using an exendable basic set of semantically and logically disambiguated (serially) nested language chunks.

> I have collaborated with some colleagues in developing some advanced AI applications.  Different users require totally different interfaces, and we accept their specifications (whatever they may be) and support whatever requirements they ask for.
 

Here I would disagree: this is not about delivering a User Interface, but an 'extended medium of language', which would have multiple User Interfaces, but, first and foremost, as a medium, must be usable and appeal to a larger audience (the particular linguistic/language community). Surface words (compounds of letters or phones) are forms of the medium language. In an age of augmentation, the medium language could be extended to better disambiguate the intentions and meanings of language (form) generators. And that is what this is about: improved artificial (artificially augmented) langauge forms (such as linear units / disambiguating language chunks). 

> See the applications in the following slides:  http://jfsowa.com/talks/cogmem.pdf  (Note that Arun Majumdar is the first aurhor.)

As mentioned before, I admire these efficient systems, because I know (from own failures) how difficult this is. 

Best, 

Lars

[Alex Shkotin]

Lars,

Very interesting! I have only three using HOL:

(I saw (((the man) on (the hill)) with (a telescope))) = the man with  a telescope
(I saw ((the man) on ((the hill) with (a telescope)))) = the hill with a telescope
(I saw ((the man) on (the hill)) with (a telescope)) = I saw … with a telescope
this last case is most interesting as "with" modifies "saw" from binary to ternary operator: X saw Y with Z

But what is the structure of the sentence itself? For the last version, I see it as

And the only addition to Chomsky context-free languages [1] is a blank node!

Alex 

[1] https://www.researchgate.net/publication/45865036_Program_structure

сб, 18 дек. 2021 г. в 00:27, Dr. Lars Ludwig <ma...@lars-ludwig.com>:

[Michael DeBellis]

I disagree that " English is normally stated in ways that are not ambiguous." The reason it seems that way is that we are so good at providing context. Consider the last sentence I wrote where I said "The reason it seems that way". If you just took that sentence on its own you have no way to tell what "it" refers to. It is impossible. (The same for "it" in the previous sentence). That's why just looking at individual sentences or even just written English rather than English as spoken isn't a true test of natural language processing. 

That brings me to my next question: Consider the following:

"The officer asked me how I could see the man on the hill when the hill was so far away. I told him I was setting up my astronomy equipment and before looking at the stars I saw the man on the hill with a telescope."

Now the syntax of "I saw the man on the hill with a telescope." should be unambiguous. I'm curious if your system can use the contextual information to give just one syntax for that sentence rather than many? 

Michael

You received this message because you are subscribed to a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/ssb8bczkoig/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/827266145.568887.1639810843932%40email.ionos.de.

[Dr. Lars Ludwig]

Dear Michael, 

Michael> I disagree that " English is normally stated in ways that are not ambiguous." 

Here you are referring to a statement that Doug made. I wonder what it means: 'Is English stated in ways that are not ambiguous' in the sense that the statement at the time of stating is not ambiguous (that is taking the context into consideration), or is it the result of the ways of stating English not ambiguous (even after the context of stating is gone). To the first I would agree, to the second I would disagree.   

Michael> "The officer asked me how I could see the man on the hill when the hill was so far away. I told him I was setting up my astronomy equipment and before looking at the stars I saw the man on the hill with a telescope." Now the syntax of "I saw the man on the hill with a telescope." should be unambiguous. I'm curious if your system can use the contextual information to give just one syntax for that sentence rather than many?

I guess you refer to the system I use. - It is not really made for displaying or ontologizing long nested propositions, but it's possible:    

The chunks/linear units nested inside this proposition are unique instances (only valid in their context). I am normaly not interested in contextually specialized (episodic) linear units due to the formal redundancy they create in the ontology. The way to disambiguate them is to tag them (in reverse nesting order) as belonging to a specific wider linear unit context (they are literals/significants of ...) by using the respetive underlying triple statement as a triple mark (TM) annotation. This allows the reuse of these components in other propositions from the same context, which also solves the problem of ambiguity of reference in formalization. The rendering above is not 100% perfect, as the system omits rendering the deepest levels of nesting. What's important, though, is to note that it is possible to construct the syntax and dependencies of (English) sentences by using linear units as ontology concepts. That's a powerful thought, as it forms a direct bridge (and potential information flow) from ontology to NLP (and language models). 

Best, 

Lars      

 

[John F Sowa]

Kingsley,

 

There is a simple reason for my claim:  No advanced AI projects use the tools that were designed for the Semantic Web in 2005 and later updates..

 

JFS:  We have more advanced tools now, but even the projects in cogmem.pdf (prior to 2010) are far more advanced than anything that can be done with the Semantic Web tools.   (See http://jfsowa.com/talks/cogmem.pdf )

KI: How could that assertion be tested objectively? For instance, are there common live examples to which technologies from both realms could be applied?  Personally, I prefer to demonstrate whatever I assert in situations like this using live examples.

As I frequently say, the world economy runs on legacy systems.  The kinds of tools that your company (and others) provide enable those systems to migrate from 20th c. batch systems to 21st c. onlline systems that keep their data in the cloud.  They also enable the programmers who were trained on the older technology to add new function with those tools.

 

As you and other Ontolog subscribers have shown, there is a solid market for businesses that can build a successful business with those tools.  The tools used for cogmem.pdf weren't designed for that purpose.

 

But note the legacy re-engineering project summarized in cogmem.pdf.  A company that wanted to do a complete overhaul of 40 years of legacy software (written in COBOL for IBM mainframes)  needed a complete analysis for all that COBOL, IBM Job Control Language,   huge volumes of SQL databases, and 40  years of English documentation (reports, manuals, specifications, emails, and comments in the code) with all the cross-references, etc

 

Accenture said that the project would require 40 people for two years to do that analysis.   But with the VivoMind system, Arun Majumdar and Andre Leclerc finished the job in 15 person weeks.

 

There is no SW technology that would be of any use to VivoMind for that project or any of the others described in cogmem.pdf

 

There is no SW technology that would be of any use to IBM for the  Deep Blue project that beat Karpov in chess, or for the Watson project that beat the Jeopardy champion.  There is no SW technology that enabled Google to beat the world champion in Go.  There is no SW technology that is of any use in any system that does machine translation or guides a car or a truck on the Interstate highways.

 

In summary, it's quite likely that your company with the technology you use (which includes SW tools) could do the projects for which the SW tools were designed better than VivoMind could do with Cognitive Memory.  But the SW tools could not begin to do the kinds of tasks in cogmem.pdf.

 

John

[Kingsley Idehen]

On 12/21/21 12:53 AM, John F Sowa wrote:

Kingsley,

Hi John,

 

There is a simple reason for my claim:  No advanced AI projects use the tools that were designed for the Semantic Web in 2005 and later updates..

I don't really know how to process that statement. By that I mean:

What's the comprehension enabling context for:

1. Advanced AI

2. Semantic Web in 2005, bearing in mind it's the year 2021

 

JFS:  We have more advanced tools now, but even the projects in cogmem.pdf (prior to 2010) are far more advanced than anything that can be done with the Semantic Web tools.   (See http://jfsowa.com/talks/cogmem.pdf )

KI: How could that assertion be tested objectively? For instance, are there common live examples to which technologies from both realms could be applied?  Personally, I prefer to demonstrate whatever I assert in situations like this using live examples.

As I frequently say, the world economy runs on legacy systems. 

Yes it does, and will do so for many years to come. That said, transparently integrating legacy systems with newer systems is something that benefits immensely from machine-readable entity relationship type semantics -- especially when woven into a Web constructed from hyperlinks.

Thus, wouldn't that provide a simple object foundation for testing your claims i.e., given a basic integration problem how do your solutions compare against those wired to operate on a Semantic Web?

The "Web" in "Semantic Web" is all about the global connectivity prowess unleashed by HTTP in regards to entity denotation. For instance, being able to share a link to a piece of novel data integration is a very simple and objective test, IMHO.

The kinds of tools that your company (and others) provide enable those systems to migrate from 20th c. batch systems to 21st c. onlline systems that keep their data in the cloud.  They also enable the programmers who were trained on the older technology to add new function with those tools.

 

As you and other Ontolog subscribers have shown, there is a solid market for businesses that can build a successful business with those tools.  The tools used for cogmem.pdf weren't designed for that purpose.

If one set of tools for legacy data integration fall into the "Semantic Web Tools" category and what's denoted by http://www.jfsowa.com/talks/cogmem.pdf#whatsDescribedInThisDoc falls into another, how can anyone perform an objective comparison?

 

But note the legacy re-engineering project summarized in cogmem.pdf.  A company that wanted to do a complete overhaul of 40 years of legacy software (written in COBOL for IBM mainframes)  needed a complete analysis for all that COBOL, IBM Job Control Language,   huge volumes of SQL databases, and 40  years of English documentation (reports, manuals, specifications, emails, and comments in the code) with all the cross-references, etc

Yes, but is that a canonical use-case for objective comparison? As you state above: "A company..." which in my worldview appears quite esoteric in regards to the kind of generalization in your comparative assertions.

 

Accenture said that the project would require 40 people for two years to do that analysis.   But with the VivoMind system, Arun Majumdar and Andre Leclerc finished the job in 15 person weeks.

Okay, but what makes that situation canonical, circa 2021?

 

There is no SW technology that would be of any use to VivoMind for that project or any of the others described in cogmem.pdf

Even if that were to be true, I don't understand how one project defines an entire software technology category.

 

There is no SW technology that would be of any use to IBM for the  Deep Blue project that beat Karpov in chess, or for the Watson project that beat the Jeopardy champion.

Bearing in mind my proximity to the IBM Jeopardy challenge, I know for sure that it wouldn't have happened without Linked Data from the DBpedia Knowledge Graph.

That project was just another example of what can be achieved via loose-coupling of critical components.

There is no SW technology that enabled Google to beat the world champion in Go.  There is no SW technology that is of any use in any system that does machine translation or guides a car or a truck on the Interstate highways.

How are any of these examples canonical, bearing in mind the topic "Different strokes for different folks" ?

A Semantic Web, as you know, is more to do with melding connectivity (via resolvable identifiers e.g., hyperlinks) with structured data representation that's informed by logic.

Today, Apple, Microsoft, Google, and many other industry behemoths all incorporate "Semantic Web" technologies and concepts into their product strategies. Here's a very simple example:

When challenged by the problem of publishing information about Products, Offers, Frequently Asked Questions, Reviews, etc., these companies now resort to the "RDF deployed via HTML" approach for Knowledge Graph publication. Why? Because it offers the following benefits:

1. Search Engine Optimization -- search engines are moving away from keyword indexing to Knowledge Graphs

2. Content Management Optimization -- new data-driven workflows that address change-sensitivity challenges (i.e., content generation is becoming more like desktop publishing and mail merge patterns of yore)

3. Content Reuse -- a hyperlink that identifies a page doubles as a data source name usable by declarative query languages such as SPARQL or SQL i.e., data is no longer confined to tables in an conventional RDBMS since a Semantic Web offers a much more powerful Giant Global Entity Relationship Graph abstraction

 

In summary, it's quite likely that your company with the technology you use (which includes SW tools) could do the projects for which the SW tools were designed better than VivoMind could do with Cognitive Memory.  But the SW tools could not begin to do the kinds of tasks in cogmem.pdf.

Maybe not, but it's hard to say for sure if we don't have a properly defined premise for objective comparison.

Links:

[1] https://twitter.com/kidehen/status/1459260324742377477 -- Twitter Thread about Apple's innovative use of Knowledge Graphs re practical Semantic Web use

[2] https://twitter.com/kidehen/status/1472954226347323392 -- recent Twitter Thread about the notion of a Semantic Web and where it stands today, in response to a recent Tim O'Reilly essay

Kingsley

 

John

--
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 on the web visit https://groups.google.com/d/msgid/ontolog-forum/36d7e72d5944437da1f5ce436f957a65%40bestweb.net.

-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software   
Home Page: http://www.openlinksw.com
Community Support: https://community.openlinksw.com
Weblogs (Blogs):
Company Blog: https://medium.com/openlink-software-blog
Virtuoso Blog: https://medium.com/virtuoso-blog
Data Access Drivers Blog: https://medium.com/openlink-odbc-jdbc-ado-net-data-access-drivers

Personal Weblogs (Blogs):
Medium Blog: https://medium.com/@kidehen
Legacy Blogs: http://www.openlinksw.com/blog/~kidehen/
              http://kidehen.blogspot.com

Profile Pages:
Pinterest: https://www.pinterest.com/kidehen/
Quora: https://www.quora.com/profile/Kingsley-Uyi-Idehen
Twitter: https://twitter.com/kidehen
Google+: https://plus.google.com/+KingsleyIdehen/about
LinkedIn: http://www.linkedin.com/in/kidehen

Web Identities (WebID):
Personal: http://kingsley.idehen.net/public_home/kidehen/profile.ttl#i
        : http://id.myopenlink.net/DAV/home/KingsleyUyiIdehen/Public/kingsley.ttl#this

[John F Sowa]

Kingsley,

 

I certainly agree with you on that point:

 

KI> transparently integrating legacy systems with newer systems is something that benefits immensely from machine-readable entity relationship type semantics -- especially when woven into a Web constructed from hyperlinks.  ....The "Web" in "Semantic Web" is all about the global connectivity prowess unleashed by HTTP in regards to entity denotation. For instance, being able to share a link to a piece of novel data integration is a very simple and objective test, IMHO.

 

Since I had been working on R & D at IBM for 30 years, I realized that Development paid for my salary, but the blue sky Research was nearer and dearer to my heart.  I managed to juggle both -- do enough D to satisfy management while going off into the blue sky R.

 

The work with VivoMind, however, enabled us to introduce some of the most advanced R into successful applications that customers were willing to pay money to get.  But those applications in http://jfsowa.com/talks/cogmem.pdf were far beyond the mainstream of the technology in 2005.  Each application was a one-of-a-kind achievement.  It was not something that could be done by the mainstream programmers and system developers.

 

Today, however, that R has matured.  And there is much more to be said.

 

John

[John Singer]

John - I googled VivoMind and didn't come up with much.  Is this technology available in some commercial form we can work with?  If not, can you recommend other options.  I'm absolutely convinced that database technology has to move more towards language as the API and storage model.  

--

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 on the web visit https://groups.google.com/d/msgid/ontolog-forum/3a0b098aaa3a48faa07793e3bc3b51f9%40bestweb.net.

[John F Sowa]

John Singer>- I googled VivoMind and didn't come up with much.  Is this technology available in some commercial form we can work with?  If not, can you recommend other options.  I'm absolutely convinced that database technology has to move more towards language as the API and storage model.  

 

The technology I was describing is stilll in the research and limited application stage (only for companies and gov't agencies with deep pockets.)  My major complaint about the DAML project is that it killed the high-end R & D that could have developed much more intelligent systems. 

 

I don't blame Tim Berners-Lee, but I do blame the decidability gang, who gave us OWL  -- a pathetically weak and wimpy dead end.  It's OK for some applications.  But it's totally useless for advanced R & D.

 

That is also my major complaint about the emphasis on deep neural nets.  They can do some things very well, but they are hopelessly bad for understanding what they "learn".. 

 

For more about these issues, see the review article http://jfsowa.com/ikl and the many references it cites

 

 

Re different stroke for different folks:  I don't blame anybody for using OWL and DNNs for applications they find useful.  But I do blame the hype artists who oversell those tools and thereby distract attention from the kinds of tools discussed in the IKL article and references.

 

John

---

 

[Michael DeBellis]

I don't blame Tim Berners-Lee, but I do blame the decidability gang, who gave us OWL  -- a pathetically weak and wimpy dead end.  It's OK for some applications.  But it's totally useless for advanced R & D.

But OWL was never meant to be for advanced R & D, at least to the best of my understanding. Correct me if I'm wrong but I thought OWL was an attempt to bring some of the research from languages like KL-One, Loom, DAML+OIL out of the lab and into the real world. I agree that for research a more powerful language would be better but in my experience with people in the corporate IT world  I can't ever remember someone complaining that OWL didn't provide enough reasoning power.  What they do often worry about is that OWL reasoning is so sophisticated that it won't scale for Big Data (knowledge graphs with billions of triples). Many vendors only support a subset of OWL  for that reason. That's also reflected in what I see in large scale applications. People often choose lower level models such as RDF/RDFS or Property Graphs because they are afraid OWL won't scale. 

Michael 

--

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 a topic in the Google Groups "ontolog-forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ontolog-forum/ssb8bczkoig/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ontolog-foru...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontolog-forum/25625102a4f74c42a3a5e7495589ba42%40bestweb.net.