Image Recognition

in Eidetic Artificial Intelligence©

“[ Neural network needs] 300 million pictures of cats to be able to say whether something is a cat, cow, or dog. Intelligence is not related to big data; it’s related to small data. If you can look at a cat, extract the principles of a cat like children do, and then forever understand what a cat is, that’s intelligence.”

Pascal Kaufmann

Actually, why children studding recognize cats faster than neural network? In other words, what is the human intelligence principal architecture?

Let’s look deeper in recognize process!

But first of all, let’s simplify the task to better understanding of principles.

Let’s take the apple. This subject is easy for recognition instead of cat. Then on that exemplar it will be easy to understand the recognition mechanic.

The brain mechanism I will describe in epistemological terms, like sort of theory of knowledge. In other words, how it looks by my mind’s eye, in self vision.

I do it aforethought, because only creating alive, native reasoning process description could provide us with base for algorithmic, including mathematic formulation, neurophysiology research, creation of artificial intelligence (strong artificial intelligence) and so on.

To solve this task, let’s begin mental experiment. Imagine we are on place of one-year child. You know about what is an apple not so long. You can’t recognize it very well.

Therefore, child has slowly recognition process, which is very good for science — we can trace it!

Generally, pedagogic science, especially of first years of life is no end of work for studding algorithms research. But for some reason computer scientists didn’t look at this area, prefer to mathematics algorithms.

Begging me pardon, let me asking, when human begin studding the difficult mathematic, algebra?

Right! — In the school, from 10–12 years old.

Wait a little, how we can study machine to think, if we don’t understand how the children studding to think in the beginning of life? If we do not understanding how the studding is possible? And we shaming of to ask ours children about it?

What if just remember how we studding personally?

I remember-imaging how it was, how I was young. I am 1 year child. Mother gives me the apple.

I am looking on the apple. I don’t know what to do with this thing. It is round [1st pattern, form]*.

The plaything could be round, ball for exemplar.

I am interesting and I take apple by hand. It colored in red-yellow unruffled [2nd pattern, color], but the ball is usually colored by different colors. Thus, it is not a ball probably.

I drop it on the floor. It is not jumping. Thus that is not a ball really!

I smell it — it is sweet [3rd pattern, smell]! Thus it is food probably. O, I know many eatable yummy round things already: orange, apple, mandarin orange, peach… I even know like them named — fruits!

Circle of searching is becoming closely. Unnecessary patterns-Eidoses is rejected. I am focusing on fruit recognition! I am pushing a little on that round fruit. It is not softly [4th taction], thus mandarin orange and peach have to be rejected too.

Orange has hard pimply peel, I remember that. Besides, orange has orange color [again pattern number 2, color], thus orange has to be rejected too.

What is remaining?

Apple!

To decrease the risk of mistake I carefully take a piece of this fruit and feeling incomparable teste [5th pattern, teste] of apple juice as a reward!

It is apple absolutely! Task of recognition is complete!

Mental experiment is finished.

As you can see, during the process of recognizing child used: the eyesight which can operate with pattern of form (pattern is Eidos actually); sense of smell (Eidos of smell); taction (Eidos of feeling); etc.

All of these 6 organs of feeling are connected with Eidoses, which are conception–representation of the world. This feature of our mentality was notice by ancient already and good described by Aristotle in “Categories”. Do you remember: category of place, category of time, category of space…?

Category of form [1st pattern] is subcategory of condition category. Categories of taste, smell, color are subcategory of quality category. Thus Aristotle considered high categories which ones are top of pyramid. In the basis of pyramid we can find the everyday occurrence categories. There is category hierarchy.

But wait a little, is the same principle of category description is using by scientist in researching?

How candidate’s dissertation requirements look like? For exemplar:

- Entry (category of cause, answering the question “what for?”);

- Research subject description (category of subject, answering the question “what?”);

- Explanation of topicality (category of time, answering the question “when?”);

- References (category of space, answering the question “where… from?”)

- Conclusion (category of quality, considering value of researching)

- etc.

The question is suggesting itself: if principle different does really exist between apple description made by child and scientific description made by serious man?

The answer is NO from point of eidetic view!

May by human possibility of studding is hiding in this feature?

Human are not increasing the knowledge value (Big Data). We are just making redesign of quantity knowledge into quality knowledge. This feature is famous low of dialectic produced by Georg Hegel.

And another case: than we need to unroll wide description from small idea, we take the idea basis, pattern and saturate them with exemplars.

Look at this: child dictionary basis is forming till 7 years of age. 90% of knowledge is collecting by children till 7 years of age. Most of brain consistence is forming and developing till 7 years of age too. And further is only insignificant improvement. Do you remember famous words of Jesuits: “Give me to teach child before 7 years of age, and then you can do with him all of you wish”.

Neural connections are only improving insignificantly till the man growing, but quantity of neurons is not growing after 7 years age.

This fact is became the trap for neural network. They decide that difficulty and versatility of mind is providing with neural connections. And the intellect is product of inside neural network self-organization. After that ideologists of this method try to create neural networks deeper and most difficult. They hope that self-studding will bear inside the deep (Deep Learning).

This theory considers intellect does not existing. Intellect is immanent property of difficult organized matter. It looks like typhoon “behavior” which could be represent like demoniacal intelligence. Any system has intellect. Than the system is complicated the humans perceive this like intelligence behavior.

May be computer neural networkers are right. But let’s consider what exist first — chicken or egg?

Did categories exist before Aristotle or Aristotle just combined existing lows of language, logic and human intelligence into categories?

Possibly growing of neural connections is consequence of patterns growing, Eidoses growing, semiotic tree of the knowledge growing?

Or one does not exist without other, but cooperating organically?

Could pattern of taste arise inside of machine if machine has no taste organ?

Could human understand how it possible to teach machine, if human has no processor inside of brain?

Could neural connection be formed if nothing exists to connect? What if instead of fanciful mix of human feelings, animal instincts, social education and strong will inside — we have picture stream from social net only?

I hope the answers for all that questions could be found with help of new science — computer eidetic.

When exact sciences go at deadlock, the integral philosophy method using by computer eidetic could become a bridge between computer science, mathematics and neuro physiology. Who knows may be this way will lead us to creation of new form of mind — eidetic artificial intelligence?

__________________________________________________________________

*Here and after: pattern, alias image, alias Eidos — the element of description of single thought, model representation and definition. Psychology named them like impressions or imprints. In ordinary life we call them ideas. This term have been borrowed from ancient Greece philosopher Plato, which considers ideas are real, material. If we agree that mind models or images could be drawing at least, we could agree that they are material.

Aristotle’s logic and “Category” doctrine have been developed from Plato’s dialectic. Mathematics and physics could not be created without logic. All our modern science and our modern world couldn’t be born without logical science. Instead of mythological form of thinking, world has been managed by new science. It was born from that seed, from the Eidos of philosophy like new machine usually beginning from the model. Thus, Eidos really closely could be understood like model.