Exploring Cognitive Technology and Conceptual Graphs with Dennis L. Thomas

Dennis L Thomas, CoFounder, CEO - has 40 years of business operations, marketing, corporate sales, and technical expertise. Thomas has more than 18 years of direct experience with knowledge system research, design, and development. He is the conceptual designer of IQxCloud, a published author, and speaker at innovation and Knowledge Management conferences.   Home Page LinkedIn

Episode Transcript

(00:12):
This is Edwin k Morris,
and you are about to embark on thenext Pioneer Knowledge Services because
you need to know a digitalresource for you to listen to folks
share their experience and knowledgearound the field of knowledge management
and nonprofit work.

(00:36):
If your company or organization wouldlike to help us continue this mission and
sponsor one of our shows, email,
b y ntk pioneer k.org.
Hello, Edwin. This is, uh, DennisThomas. I'm, uh, located in Marine City,

(01:00):
Michigan. This is kind of a unique area.
It's actually one of theresort areas of Michigan.
We're located right onthe St. Clair River,
which has the distinctionof being a blue water river.
Most rivers are muddy and they're kindof brownish in, in color, but the St.
Clair River actually is very blue.
It separates the state ofMichigan from Canada and the river

(01:24):
runs from North side, whichis, uh, lake Huron down to, uh,
lake St. Clair, and then down toDetroit and to, uh, lake Erie.
So it's a very unique and special area.
I'm committed personally to usingtechnology to educate people
for teaching them readinessskills to get into business,

(01:45):
but also using it to capture thecognitive knowledge of organizations
and transferring thatknowledge to knowledge workers.
I'm the CEO of IQStrategics, which is a, uh,
cognitive software development company.
The last book that I read that I foundreally interesting was Real Teams
Win, and it's by Ted Steadingand he's outta Silicon Valley,

(02:09):
and it's all about what does ittake to put a real team together,
to have a mastermind,
to create a technology or createany kind of a business to,
and make it successful.
The topics that really interest meare cognitive technologies and how
they're different from dataand digital technologies,

(02:30):
and also AI technologies.
A lot of that resulted and came aboutbecause my business relationship with
a fellow by the name ofRichard Al Ballard, and, uh,
Ballard is a Berkeleyexperimental physicist.
He was the 110 employeewith Apple computer.
So it was an extraordinaryopportunity for me to work with him.

(02:52):
So that pretty much summarizeswho I am and what I'm doing.
You talk about technology,
but we're basically talking aboutcomputer technology, is that correct?
Yes, it is.
Okay. So what exactlyare we talking about?
Are we talking more thanjust term recognition or
key, I guess just keyword search.I mean, we're beyond that.

(03:15):
We're way beyond that.
That's correct. Yep, absolutely. All.
Right. Well explain how and why. Whwhere did we go and how did we get here?
You know, there have been many generationsof technology that have occurred,
and of course in the nineties iswhen databases and process modeling
became very important. And KMactually came into being in the,

(03:36):
the mid nineties in a,in a very large way,
when in two thousands, uh,Tim Burners Lee and his group,
a W three C consortium, theycame up with the idea of, uh,
XML and extensible languageand new technology called
triples. At that time,

(03:57):
the whole idea of capturing knowledgeand representing knowledge became
foremost in the technologist's minds.
And then it evolved into the nextlevel where we had like conceptual
graphs, and that's kind of what a lotof technologies are doing right now.
And then we're evolving intoadvanced AI technologies.

(04:18):
What's conceptual graphs? Canyou explain to me what that is?
This is part of the differencebetween us and those technologies.
A conceptual graph basically uses terms.
Let's say you have the city calledWashington, DC mm-hmm. ,
and then a conceptual graph, well,
what's related to Washington dc So therewould be a line coming down with that

(04:40):
mm-hmm. , and maybeit'll say labeled belongs to,
and then it'll say United Statesof America. Mm-hmm. .
So the conceptual graph is you havea city and then it says belongs to,
and then you have theUnited States of America.
Uh, just in terminology speak is thatwould be an interpretation of an ontology.
It is. We're going and we're actuallydoing is making a better user interface,

(05:04):
I would think when you can start tosee connections and the why and what
and how things are related.
That's true. The, the conceptual graphs.
And so they're basically a backendtechnology. So as a backend technology,
it's not something that a personnormally is going to see. Oh,
usually what they'll do,
and this is probably the part of thecomplexity of common technology or regular

(05:28):
data technology, is they have taxonomies,
and then those taxonomieshave to be translated,
be mapped to a user taxonomy,
and then that user taxonomy istranslated and presented through a
front end technology.
And that front end technology iswhat individuals interact with.

(05:49):
Well, it sounds like there's a lotof bridges connecting all this.
There are many bridges connecting .
So depending on who yourgatekeeper is at the toll booth,
you may not get all the data or,or there's a good opportunity,
something went awry andwent to the wrong place.
How do you map all thisstuff and keep it straight?

(06:10):
Well, number one, it's very complex.
It requires highlyskilled people to do it,
and then a lot of testing to do it, andthere's a huge amount of costs. Okay.
There are tens of billions of dollarsthat have put into developing this
technology and trying to makeit work. For example, um,
project managers with the, uh,D O D or with the government,

(06:34):
and they had projects and it maybetook them two years to do that.
And what they came out with was like aroom full of banker boxes and that's it,
that, that ended there. Actually,
most of these technologies upto 65% of them or more actually
fail. And so they're, that's theproblem with the technology. Wow.

(06:54):
It's the complex.
We make the leap, I'll say,
from ontological connections to a more
sophisticated methodology inwhich machine learning takes
a new spin, the machine learning advance.
And how long has that been in play?Maybe 20 years machine learning,

(07:15):
or is it longer.
Than now? Well, probably that long,but really intensely since, uh, 2010.
What was the switch point in 2010?
So in 2010 in particular,
AI took a major jump,
and part of the majorjump is neural networks.
And that's a technology, it's an,an older technology, but you know,

(07:35):
as technology advances,
older technologies get new life andare reapplied, and in this case,
neural networks.
What they're designed to do is theysit between an application and the
back end of a, a technology, a server,
and what they do is they capturethe terminology that's being

(07:56):
used through that application,whatever it may be.
It also identifies thepatterns of thought,
and it turns out that people thinkin terms of patterns of thought,
but data patterns or thought aredifferent from the cognitive human
patterns of thought, but generaltechnology doesn't recognize that.

(08:18):
And so, uh,
what happens with the neural networksis then all of a sudden they started
building what, you know,
the new things that are now calledlarge language models, for example,
chat, G P T mm-hmm. .
And so they have this massivestores of all of these phrases
that are used across allindustries and all nations,

(08:39):
wherever it may be massive amountsof these phrases, language phrases,
and they call patterns of thought.
And then all of this differenttechnology terminology.
And then they piece thetechnology, the, uh, chat,
g p t structures puts this all intogether mm-hmm. , and it's,
that process is called generative ai.

(09:01):
So it's a generative AIprocess of putting all of this
data together and then spittingout the answer to the question,
whatever the answers may be. Uh,
Chan g p t actually is anincredibly honest technology,
and it'll tell you what its limitationsor faults are if you ask for them. And,

(09:21):
uh, there are, there are many .
Well, wow. We, we've covered, uh,
quite a fast train of revolutionand evolution of the digital
space.
Where does cognition startfor real in software?
That is an absolute fundamentalquestion. The thing with software,

(09:42):
and this is my whole thing, we'vedeveloped a knowledge science.
Knowledge science is knowledge,e school's, cognition plus data,
which is facts, and theninformation about those facts.
It's a cognitive process in almosteverything that exists within the world,
including what's occurring inour brain is a cognitive process.

(10:05):
But computers are based on logic.
And logic requires selfconsistency and of course language
itself primarily inventedpretty much in India,
but definitely established myAristotle. Logic is used in language.
So language is structured accordingto logic as well. In databases,

(10:26):
mathematics is all about logic.
And so when they structure databases andbuild algorithms, they require logic.
And logic requires self consistency.
So therefore any kind of a dataproduct that's created has to
operate, uh, according to machine logic,
and it requires self consistency,

(10:47):
and it'll always look forthe self-consistent answer.
Computers are diametricallyopposed to humans.
Humans are forced to operate the way thetechnologies demand that they interact
with the technology,
whereas cognitive technologies arehuman oriented mm-hmm. ,
and they work the way thatpeople naturally think,

(11:08):
making them a proper andacceptable way for humans to
learn. And that's what ourtechnology is designed to do.
Miriam Webster defines logic asa science that deals with the
principles and criteria of validity,
of inference and demonstration,
the science of the formal principlesof reasoning, as we all know,

(11:32):
not everybody reasons the sameway across the human spectrum.
How do you protect reasoning from bias?
I, I, I think a more fundamentalquestion at this point. How, how,
what is the difference betweenmachine logic and human cognitive
thinking? Okay.
All right. All right,let's go down that road.

(11:53):
The difference between human cognitivethinking and computers is that
cognitive thinking,
rational thinking basically is highlyimaginative. Hmm. As human beings,
we're very goal-oriented because ourbrains are constantly asking questions,
where are we? Is there a threat in thisenvironment? Mm-hmm. ,

(12:14):
what, what does this situation mean to me?
Are there values to the situationwith me? And so the human, cognitive,
rational thinking isall about understanding situations and circumstances.
Mm-hmm. data systemsare all about trying to represent
situations and circumstances,
but they can't really do itbecause it's all logic based.

(12:39):
Mm-hmm. ,
how does a human being think the humanbeing thinks in terms of common patterns
of thought? And this occurs acrossall people regardless of nationality,
language or culture. And how do wethink, we think in terms of compositions,
composition is an, and this construction.
So when we walk into BestBuy and we look around,

(13:02):
what we see is a compositionof departments. You know,
you could have the TV department,the appliance department,
the cell phone department,computer department, TV department.
When we go back andtalk to the salesperson,
then we ask them a questionand let's say, well,
what kind of TVs do you have? Theyautomatically go into a taxonomic mode,

(13:24):
they think in terms oftaxonomies. Mm-hmm. ,
that's where you have these linearstructures. Mm-hmm. ,
and it's an and or structure. So wehave this kind of tv, that kind of tv,
and it goes all the way through thosekind of thought patterns, like a sequence,
well how do I do this?
When you ask a how question automaticallyyou're into a sequence pattern of

(13:45):
thought.
And this is a cognitive pattern thoughtthat exists across all nations and all
people. Mm-hmm. . Sothose are common patterns of thought,
and that's what our technologyis designed to represent.
And it sounds like a whirlwind ofcontinual change as advances happen
all across the spectrum. Uh,
and I can't imagine what it's liketrying to just keep up in the world of

(14:08):
software development as youare a proponent and a longtime
kmr. Can you define for mewhat knowledge management is?
Well,
knowledge management is the artand science of identifying where
the knowledge exists within anorganization and how it flows through that
organization. Mm-hmm. ,
and how to capture that knowledgeand present it to represent

(14:32):
that knowledge to end users. And I'llgive you a classic example. Okay.
And this is a major trend that'staking place right now within industry.
And I'm in Michigan and we have lotsof manufacturing companies here,
and there's a major new wavetaking place within the industry.
The problem is,
is they have all of theend-to-end business procedures or

(14:55):
end-to-end business processes, uh,organized within e r P systems,
enterprise, uh, resource managementsystems. It's the backend of technology.
They're all hidden in blackboxes, so nobody can see that.
But the machine works with theseend-to-end business processes.
So the trend right now is how doyou get these end-to-end business

(15:18):
processes? How do you representthem within a technology,
a knowledge technology,
a KM technology so that endusers can actually see them
and understand what's going onand be able to connect the dots.
That is what the KM role is right now.
I appreciate your timeand your expertise today.

(15:39):
It was a very interestingjourney in the digital world.
Thank you very much. Have a good.
Day. Thank you, Dennis.And thanks for being here.
You have just finished our latest becauseyou need to know a public service of
Pioneer Knowledge Services.
Please join us on LinkedInand find [e-mail].