The immense power of human intelligence

[John F Sowa]

The article on phaneroscopy, which I have finally finished, shows the immense power of human intelligence, of which LLMs can simulate only a tiny aspect.  Compared to earlier work on machine translation, it's an important aspect.  But in a comparison with what the human brain can do, it's pathetically weak.  In fact, it's pathetically weak compared to a rat brain.

The concluding Section 7, which is attached below, shows an illustration of an intelligent human system (Figure 18), a design for an intelligent AI system (Figure 19), and a design for AI systems that implement aspects of intelligence for practical computer systems.  They aren't AGI systems, but they would have a chance of supporting  systems that are more powerful than anything done with LLMs by themselves.  In fact, they could use LLMs to support a language interface.

The preceding Section 6 surveyed aspects of human intelligence.  But the references that follow Section 7 include two slide sets for talks that survey aspects of human intelligence and contain many references to the original research reports.  To begin, see https://jfsowa.com/talks/natlog.pdf .  For more, see https://jfsowa.com/talks/vrmind.pdf 

As I have said many times, I fully appreciate the value of LLM technology.  But ongoing research in neuroscience shows that animal brains, from the rat on up, are vastly more powerful.   AGI is far in the future.  As  I would bet, not in the 21st century.

John

 

[David Poole]

What is missing from this (and most of the work on LLMs) is the importance of values/preferences. This should be familiar with anyone what has read Alice in Wonderland:

Alice … went on “Would you please tell me, please, which way I ought to go from here?”
“That depends a good deal on where you want to get to,” said the Cat.
“I don’t much care where –” said Alice.
“Then it doesn’t matter which way you go,” said the Cat.
– Lewis Carroll (1832–1898)
Alice’s Adventures in Wonderland, 1865

You can’t learn what is right and wrong from observing the world. (Imagine a computer trying to trying to learn “thou shall not kill” from observing human behavior). It has to come from outside. (Arguably you can get it from evolution; societies that have poor morals will die out).

This is a main theme of our AI textbook. See https://artint.info/3e/html/ArtInt3e.Ch1.S3.html <seehttps://artint.info/3e/html/ArtInt3e.Ch1.S3.html> or https://artint.info/3e/html/ArtInt3e.Ch12.S1.html#SS3.p7 for example.

You need to have values/preferences as an input to the point where actions are chosen. I don’t disagree with any other part.

David



> On Jan 29, 2024, at 3:36 PM, John F Sowa <so...@bestweb.net> wrote:
>
> [CAUTION: Non-UBC Email]

> --
> 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 email, see http://ontologforum.org/info/
> ---
> You received this message because you are subscribed to the Google Groups "ontology-summit" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/ontology-summit/22c2e2683acb46ef8ef91ab49fa9d89a%40bestweb.net.
> <Section7.pdf>

[Alex Shkotin]

John,

Congrats! Where is it possible to read your article in full?

I like this "an intelligent agent — human, animal, or robot" as we keep our eye on progress in robotics.

And following David's direction: the central point of OODA loop should be THE GOAL, as we are interested mostly about the goal an intelligent agent is acting for.

Alex

вт, 30 янв. 2024 г. в 02:36, John F Sowa <so...@bestweb.net>:

--

[Pushkar Sawant]

Dr. John Sowa,

I refer to one point in the writeup - that "Logic is not based on psychology, but psychology is based on logic." Althought this looks true, I can show that numbers and hence Mathematics are based uppon Psychology. Ultimately, the perception and understanding of everything is in the mind. (I am not so sure about logic).

Here is the argument - 

Why cant some things be measured?

Why cant some things be measured and we have to remain resorted to 'qualitativeness' for them? Consider a grain of sugar. It has properties - like, it is one, it is sweet, it is white etc. What is it in the property 'one' of sugar and not in the other properties like 'sweet', 'white' etc. that from the former, a Mathematics can be made and from the latter the same cannot be made? So let's see what these "numbers" are?

How do we teach numbers? We show one stick, then a stick and the same stick together, then a stick and a stick and a stick together, and so on. The first we give a NAME as "one (1)", the second as "two (2)", the third as "three (3)" and so on. So for us, the different numbers -  1, 2, 3 etc. are perceived as the different "togethernesses"/"collectivenesses" of the same sticks (or same balls or same of anything else). So to understand what numbers are, we essentially need to only understand what the first stick or the number "1" is (since each number is about progressively adding 1 to the previous one and so on, till you reach 1). What is 1? Here is my definition - "There is 1 of X if X is what I am considering at an instance". (I can also do this consideration for more duration than an instance, but the fundamental necessity is that it should be considered for an instance at least). So there is one glass, if I am considering that glass at an instance (or for more time). There is 1 arbitrary-shaped section/part of a glass defined by an imaginary boundary if I am considering that part/section at an instance.

Let's look into what this "consideration (by the mind)", of what we call numbers above, is. I am perceiving the togetherness of the sticks and cognizing the same as a certain unique number. How do we perceive this togetherness? By our visual senses (eyes). But wait, this seeing is constituted of memory of the mind, since, as I move my eyes over the sticks'-spread I use the knowledge put in memory of what I have seen just before where I am at present, while scanning the sticks'-spread. So there is this vision-memory intertwining which constitutes this consideration by the mind. (Minsky also says in The Society of Mind that "vision is intertwined with memory"). There is more. I can also imagine ONE glass in my mind at an instance. So this "considering" is composed of the triplet of ((vision+memory) + imagination). Hence whatever satisfies this triplet can be ascribed the number 1, or numbers in general, and hence be measured.

There cannot be "1 sweetness" because the perceptive experience of sweetness doesn't satisfy the above triplet. (Note - color falls somewhere between measurable and non-measurable because the 'memory' part of the triplet (while scanning a color) isn't quite of discrete memory bits, but is rather "continuous"). 

Hence the immeasurability of some properties (qualities) is not due to those qualities themselves, nor due to the human mental machinery, but due to the limited way in which some things called "NUMBERS" themselves have been defined in the very first place. So, for "measuring" "immeasurable" qualities we need to invent something broader and more encapsulating than the present idea of "numbers", and related to our Cognitive faculties.

Sincerely,

Pushkar.

[Alex Shkotin]

Hi Eric,

We are here and at the ontolog-forum google group to support https://ontologforum.com/index.php/WikiHomePage and discuss formal ontology mostly :-)

What direction of R&D is yours?

Alex

ср, 31 янв. 2024 г. в 15:53, Eric Lindahl <eric.l...@gmail.com>:

Hi,

Is the Google group the only forum for discussion?

Ordered sets of shortest equivalent multimodal descriptions is fairly close to Peirce’s semiotics.

Something neural encodings of multi resolution  spherical harmonics can do, and where some multimodal LLM are heading.

Thanks,

Eric

--

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 email, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontology-summit" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ontology-summit/22c2e2683acb46ef8ef91ab49fa9d89a%40bestweb.net.

--
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 email, see http://ontologforum.org/info/
---
You received this message because you are subscribed to the Google Groups "ontology-summit" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ontology-summ...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ontology-summit/CAGv33PtdTZim9su%3Driopwj5cpVjq9yBOBPH7PqtrU9N9OoMcbA%40mail.gmail.com.