The hope of the researchers trying to create artificial intelligence is that "it will be possible to turn the computer from just a fast calculating machine into a device that will imitate the operation of the human brain.
This article was first published in "People and Computers" issue no. 54 1987 .15-3
0.4
When a group of people express an opinion about a so-and-so that he is an intelligent person, it must be assumed that the intention of the members of the group is the same. On the other hand, the dictionary gives several definitions for the word intelligence. There are three definitions in the Ibn Shoshan dictionary: a. Education, B. Development, C. opinion. In the Oxford dictionary there are two definitions:
1. The power to think, learning, understanding, knowing-mental ability
2. News, information with references to important events
Each definition can stand on its own and all definitions can exist simultaneously. In an attempt to put all the definitions under one roof, we will formulate a super definition and say that intelligence is a dynamic process of acquiring knowledge, understanding and reacting with reference to the events that occur in the environment of the acquirer of the knowledge. This process is essentially endless. The hope of the researchers trying to create artificial intelligence is that "it will be possible to turn the computer from just a fast calculating machine into a device that will imitate the operation of the human brain. In this way, the computer will be able to perform diverse thinking operations, draw conclusions and even learn" (1). In accordance with the super definition that we formulated, the hope of the researchers is to develop a machine that knows and will be infinitely capable of acquiring knowledge, understanding and reacting with reference to the events taking place in its environment. is that so?
Its structure, the way the brain works and the ways of thinking are still shrouded in fog. Although certain areas of the brain and their ways of working and the structure of logical thinking are known, what are the structures of intuitive thinking, emotional thinking and extrasensory perception? These questions still have no answers. Here a fundamental question arises. How can you build a machine that imitates the human brain, when it is still not well known? Although the computer is capable of performing a number of human actions and even better than the human - great memory capacity and extremely high speed of action, but another question arises - can the machine that has been embedded with several mental qualities be called a brain? It is even possible that trying to build a thinking machine while the brain is still unknown will lead to unexpected results.
Man derives his knowledge, understanding and reactions from information that reaches his mind through the five senses, i.e. sight, hearing, smell, touch and taste. The information about a certain environmental item can be received through one of the senses, several senses or all the senses together. When a so-and-so talks to an unknown, the so-and-so receives his information about the unknown through the visual input system and through the voice input system. The information that flows into the mind of a certain person through these two input systems, allows him to know the unknown. This is an as-is situation. What happens when, for whatever reason, the information coming from one of these systems does not complement the information coming from the other input system, but contradicts it? A famous example of this is the biblical sentence "The voice, the voice of Jacob and the hands of Esau." There is no Yitzhak, able in this situation to know who is who. What types of thinking are needed to solve the problem - logical thinking, value thinking, experience? When value judgment or experience is required, the computer will be useless. It's easier when the knowledge requires the activation of more than two senses or the activation of all the senses together.
At the base of human language are the vowels and the letters (except the pictorial scripts such as the Chinese and Japanese script). The vowels are pronounced by the vocal output system - the vocal cords and received by the vocal input system - the sense of hearing. The letters are pronounced by the visual output system - the writing hand and are absorbed by the visual input system - the sense of sight. The total number of signs needed to establish contact between people is dozens and this includes letters, numbers and various syntax signs. In Hebrew, the scoring must also be taken into account. Words and sentences are built from the various letters and syntax signs, and mathematical formulas are built from appropriate numbers and syntax signs. The computer language is based on two symbols "1" and "0", from which sentences and mathematical formulas are built. Apparently computer language is in an advantageous position compared to human language. is a more economical language. We will compare the formal representation of numbers in human language - the decimal system to their representation in computer language - the binary system:
(2) 0010 (10) = 2
33 (10) =1000001 (2)
425 (10) = (0100X1010 +0010)X1010+0101 (2)
A conclusion is required that the larger the value of a number, the more signs are needed for the formal representation in a computer language, which rightly raises the question of the economical advantage of the binary system over the decimal system. Dealing with this question requires the development of new concepts:
A. Organic calculation system - calculation system of an intelligent creature.
B. Artificial calculation system - a calculation system developed by the intelligent creature.
The importance of the binary method lies in the high speeds it enables in artificial calculation systems. The organic calculation system of the intelligent creature must invest a lot of energy in additional calculation processes that are essential for its physical and probably also spiritual existence, and they include value thinking, accumulating experience, accumulating experiences, accumulating emotional charges, initiating actions and exchange - all of these are extremely essential in giving meaning to one's life the physical In order to create communication whether it is vocal or visual - i.e. writing - the intelligent creature developed thinking in symbols. When he says, for example, the number 425, each sign of the three signs indicates a certain quantitative value. Its position symbolizes the multiplication of this value by a certain factor and the combination of the three signs in this way and not otherwise (like 452) symbolizes a certain size. Writing three signs instead of over thirty signs (each sign of the mathematical syntax signs has its own formal instance in the binary method and therefore 254 more signs in this method), is much faster and more efficient, not to mention the large number of signs needed to remember, write and say any word. An obvious conclusion is that there is room in the memory for other things such as the accumulation of experience, experiences and emotional charges.
Thinking in symbols makes it possible to invest the energy instead in high calculation speeds, in memory arrays for absorbing information and knowledge and in other forms of thinking such as value thinking, associative thinking that will be discussed later, thinking that stems from experience, etc. In order for the computer to be intelligent, it must think in symbols, which raises another question and that is, is thinking in symbols an evolutionary process that can only develop in organic calculation systems or can it be developed artificially? We must not forget that the other methods of thinking are fundamentally qualitative and not quantitative. It is hard to see how quality methods can even be embedded in a computer.
Structurally, the software is a set of action instructions that follow one after the other in a conditional relationship. Each action is a condition for the next action after it: "If 'A' go to 180", "If 'B' then go to 500", "Agreement of 'A' and 'C'" etc. At the base of all software is the wired thinking. "If 'A1' then 'B1', if 'A2 then 'B2' and so on. Due to the conditioning and linear thinking, the computer language should be treated as a computer language. In a language of this type, every statement in it must be extremely detailed. We must walk through every description in it, step by step, when each step is unambiguous within us. Therefore a structural computer language (as well as its content) is a closed language. In an attempt to make the computer language more flexible, various ideas are tested:
A. Development of computers for parallel processing which decompose the problem presented to them into factors and handle them simultaneously similar to the activity of the human brain (1).
B. Development of software according to which the information stored in the computer is sorted and different methods of operation are defined. For example, when building a medical specialist computer, the following actions are performed:
1. The computer is fed with the data of a given disease.
2. At the same time the computer produces additional data related to this disease.
3. The computer searches for an alternative solution suitable for the new data set (the disease data and the additional data)
4. This operation is repeated several times until a final answer is obtained (1).
third. Development of software that allows the computer to develop its own theories (1).
D. Development of programming languages in which the statement is less detailed and each sentence includes more than one item of information. For example, the sentence (husband XY), means that if the husband is (XY), then (wife YX). If (brother ZY ), then (brother YZ ) (2). Such a language is inherently more economical and less cumbersome. In this language, the interrelationships between the facts can change depending on the use of the knowledge bases.
There is no doubt that these flexibilities are of great importance in the development and refinement of computers and computer languages. The question is how much these flexibilities can contribute to artificial intelligence. In his thinking processes, the owner of the brain can think both linearly and exponentially. The characteristics of exponential thinking is that it does not have to be done gradually, but in unexpected jumps and directions. On the computer as mentioned, the progress is step by step "if a1 then b1", "if a2 then b2", "if a3 then b3" until you reach the last requested step for example "if a30 then b30". In exponential thinking it is possible to reach almost up to "B30". It is also possible to reach unexpected directions, and this is based on the assumption that "if A, then H, or if A, then also F, and so on." The origin of turning to the unexpected directions can in many cases be in associative thinking - thinking that finds a certain connection between similar phenomena or so-called that there is no connection between them and no, after all, many of the scientific discoveries originate from exponential or associative thinking. These ways of thinking stem from some sudden insight. It is interesting, for example, to check how long it would have taken a computer to invent the telephone, or compare it to develop the theory of relativity. An obvious conclusion is that both exponential thinking and associative thinking are more effective than linear thinking. The computer is not capable of shouting "Eureka".
One of the reasons for the flexibility of human language is the different meanings that can be derived from a given word or any sentence. Three types of meanings (possibly more) will be identified here to illustrate the capabilities inherent in the language.
A. A single word has one or more meanings, each meaning gives a different content to the sentence in which this word appears. the trial
He is at the bank now can be interpreted as "he is at the bank now" or "he is at the bank of the river now." The Hebrew sentence "the yeshiva was difficult" can also be interpreted in two ways. One that the very act of sitting on a chair involved discomfort due to its structure, or that it was a gathering of a number of people for a specific purpose and that was conducted in a difficult manner. The meaning depends on the context in which it is given.
B. A given sentence has a meaning that differs from the meaning of the individual words that make it up. The sentence from sight far from heart has a different meaning than the meaning of the individual words that make it up. A well-known joke about this sentence says that a computer was taught English and Russian. For his test we gave him an instruction to translate this sentence from English to Russian and then to translate it back to English. The result obtained is idiot invisible.
third. Changing the place of words in a sentence for the purpose of emphasizing certain things. In the sentence "Reuven went to the house" the emphasis is on Reuven and in the sentence "to the house Reuven went" the emphasis is on "the house".
It is the ability to vary the meanings in sentences and statements in human language that gives it its nuances, subtleties, liveliness and beauty that cannot exist in a language where every statement is unequivocal. These meanings carry on their backs the experience, the beings, the experiences and the emotional sentences and in these the essence of the knowledge is embodied.
The ability of the spoken word to convey meanings and messages that the same written word cannot convey. This ability lies in the intonations used while speaking. They can convey feelings of admiration, wonder, fear, horror, aggression, disdain, etc. and can also testify to the character of the speaker. Qualities that can betray self-confidence, arrogance, begging, indifference, etc. The intonations depend on the will of the speaker. When the speaker wants to instill an atmosphere of security, for example, he will adapt his voice to the circumstances in which this atmosphere is needed, and on the other hand, a person who wants to prostrate himself before his superior would also know how to pray with his voice. After all, in order to change intonations with the intention of conveying messages, a great deal of knowledge is required. Is the computer capable of this?
Arthur C. Clark (3) proposed that intelligence is the product of a minimal number of nerve cells connected to each other, where this number is the same as the number of nerve cells found in the human brain. This number is defined as critical mass. According to this idea, as soon as several computers are connected whose sum of switching cells is the same as the critical mass, a creature whose brain is identical to the human brain will be obtained. Which raises two questions. One that was mentioned is whether it is even possible to imitate the human brain, while this is still unknown and the second question, is the critical mass enough to make the brain to which it belongs intelligent?
If we disconnect the computer from its power source it will not respond at all, a behavior not found even in animals. They will realize that they are in danger. They will either run away or fight for their lives, which the computer does not do at all. In this sense it can be said with certainty that the animal from the computer is allowed.
The attempts to develop artificial intelligence, their contribution is important in the development of computers and computer languages. However, we must not forget that this issue is extremely problematic. Most likely, a machine will never be built to imitate the human brain. The complexity of the brain is too high to be able to summarize its operation in logical thinking alone. There are essentially other forms of thinking that are the ones that prove to be the most effective and paradoxically the proof of their effectiveness can be done by using energetic considerations.
Sources
1. "Artificial intelligence - a new computing era" computers no. November 18 - December 1982 p. 106-108
2. "Artificial intelligence - the gospel of the fifth generation" Computers No. March 33, 1984 p. 22-24
3. Arthur C. Clerk: Dial M and you will receive Morgenstein" The wind blowing from the sun, published by Masada Ramat Gan 1978 229 p.
4. David Ferris - "Artificial Intelligence - New Era or Illusion" Computers no. January 19, 1983 p. 16
5. Gad Winter - "Artificial Intelligence" Computers no. July 25, 1983 p. 38-39
6. Yeshayahu Leibovitz - "Artificial intelligence - really?" computers no. March 33, 1984 p. 18-20
7. David Skelly - "Brain-like Artificial Intelligence" Computers no. March 33, 1984 p. 62-64
. Mitchell Waldrop- “Machination of Mind” Science 85 March pp. 38-45 8
9. Rose F.-"The black knight of AI" Science 85 pp. 46-51
29 תגובות
Elad,
I agree with what you say but in my opinion the video missed a few points although the general idea is correct. Tyson talks as if a 2% difference in DNA means there is a 2% difference in the brain (he said we are 2% smarter than chimpanzees) which is simply not true. The cheetah is genetically close to the cat, but this closeness does not indicate a closeness in their running speed. Not to mention that DNA is not some blueprint of the human body, so even a small difference between proteins that are active in brain development can lead to a huge difference in the final product. In any case, this is all about the video and not what you said.
safkan,
I agree that excessive importance should not be attached to the mirror experiment, but it should not be canceled either, doesn't it surprise you that certain animals react differently to their reflection than, say, to a recorded film that shows them in the past?
The joke is familiar but in my opinion is not appropriate here, the experiment with the mirror (which by the way was mentioned by another commenter) in my opinion is a pretty good indication of awareness.
By the way, how can you check if the person standing next to you has self-awareness? True, he will tell you "I am aware of myself!" But a computer program can also say that.
Elad
Self-awareness is a complex and vague quality that cannot be measured by testing "filters" such as the animal's reactions to its appearance in the mirror. These filters are limited to the animal's cognitive ability, so even if an animal fails to pass a test through a cognitive filter, it does not mean that it does not have self-awareness.
Excessive importance is attached to testing with a filter in the famous joke about a scientist who tested a flea's sense of hearing by cutting off its legs.
The joke about the scientist and the flea's sense of hearing.
============================
An important scientist cut off one leg of a flea and then shouted "jump" - and the flea jumped.
Then the scientist cut off the second leg of the flea and shouted "jump" - and the flea jumped.
Then the scientist cut off a third leg of the flea and shouted "jump" - and the flea jumped.
Then the scientist cut off the fourth leg of the flea and shouted "jump" - and the flea jumped.
Then the scientist cut off a fifth leg of the flea and shouted "jump" - and the flea jumped.
Then the scientist cut off the sixth (last) leg of the flea and shouted "jump" - this time the flea *did not jump*.
The famous scientist published an article in NATURE that said:
When you cut off all the legs of a flea, it loses its sense of hearing.
Correction - "the one we know".
Zvi, you are right, as defined by a famous neuroscientist whose name I forgot, "self-awareness" is like a lamp operated by a dimmer, its light can be weak, or very strong (bright) and any value in between. Dolphins, elephants and monkeys also have self-awareness, but it is probably of a low intensity, we have a self-awareness that is high intensity, with us the light is "brighter". It is very possible that in the future we will be able to build computers/robots that will have an even higher level of self-awareness than the one we know.
See here:
http://www.youtube.com/watch?v=uUmfw_lJO0s
Elad,
In such a case, the question arises as to what exactly self-awareness is. There are animals that have managed to recognize themselves in the mirror (dolphins, apes, elephants...) so it might be correct to say that our self-awareness is more developed than theirs, but not something that exists only with us.
By the way, I think that the most complex (and impressive!) ability of our brain is self-awareness and not emotions... As someone who raised animals in his day, I can tell you for sure that dogs and cats also have emotions and they also have moods.... They too can be happy or sad.
What really sets us apart is self-awareness.
In the closet you are wrong, there are parts of the brain that there is no need to imitate in the simulation, for example life support systems such as blood vessels whose purpose is to bring food and oxygen supply to the brain cells, there is no reason to include this in the simulation.
Even when we imitated the flight action of the birds we did not have to build airplanes with feathers or imitate the flapping action of the wings for our aircraft to fly in the air.
Many times when trying to imitate a very complex system it is enough to introduce the basic principles into the model and there is no need to make a one-to-one copy of the complex system down to the level of the single molecule or atom.
Human brains are also not the same, each person has a brain that has a unique pattern to them, there are almost 7 billion people in our world, no one has a brain that is identical "down to the level of the single molecule" to another person's brain, but despite this, everyone thinks, and everyone have self-awareness and feelings.
That is, as you can see, your claim is simply incorrect, and it is enough that the basic pattern is correct and everything else will already be created by itself (awareness, thoughts...).
Thanks for the correction, but usually it is customary to write the word without an A (Baron Rothschild) otherwise it just causes confusion....
Sabdarmish Yehuda!
Are you saying that the brain has redundant parts? As long as this is not the case, in order to reach the most difficult and complex abilities of the brain, which are undoubtedly the emotions, you need at least an exact imitation of the brain without missing a single molecule.
When the computer shouts "Save me, save me, don't disconnect me" I will understand that it was configured that way in its software, and that's all.
to someone!
It is also possible to build an electronic imitation of the biological structure of the brain. It's just a lot more complicated than building it like the original.
I think you are misreading the name: baron and not ba-aron. Baron is a title of nobility.
someone
Do you believe that a computer + software will be able to imitate an organism as long as they know how to define the organism itself?
It is as you say, that a tower that is completely identical to the Eiffel Tower can be built, solely by looking at a photograph of it.
to Barron
You said: "For the computer to really feel, you need to build an exact imitation of the brain on all its neurons" end of quote.
Well, defining a completely unproven definition, who told you it should be like that?, that's your opinion, and maybe the opinion of others, but it's not a binding and proven statement.
When a computer screams, "Please, please, don't disconnect me from the electricity, I want to live"
What will you think then??
(:))
good week
Sabdarmish Yehuda
In the closet, an imitation of biological systems can also be created on a computer as we created simulations of many other physical systems on a computer.
As for the nickname you chose, usually that term has some folk meaning that I don't know if you meant.
point!
You don't need intelligence to play chess. Even a calculator knows how to calculate better than the best mathematician.
to someone!
In order for the computer to really feel, you have to build an exact imitation of the brain on all its neurons. That is, not an electronic computer but a biological computer. Or in other words to build a human brain. You don't really think they would ever be able to do such a thing.
What is so special about the name I chose to write under?
תגובה
The computer is certainly day by day more and more successful in performing actions better than the human, but those actions did not go beyond the field I defined. A field that is affected by unique factors for each person in a different and unique way, heredity, character, education, environment, health and many more. A non-measurable field is not,
We must not forget that we are an organism also made up of viruses, bacteria and types of parasites, not some independent single creature
Seriously, what's really ridiculous is thinking that we won't be able to build a computerized system that thinks and acts like a human, as if we have some hidden "soul" or some magic powder that makes us develop self-awareness, the ability to think and our other intellectual abilities.
Today there is automatic facial recognition in every camera, we once claimed that a computer would never be able to deal with such challenges that are "exclusive" to humans, we were told that a computer would not be able to beat world champions in a game of chess, would not be able to drive a car, fly a plane and a long list of other tasks that today Computers are able to perform and even better than humans.
In 30 years we will have computers that think and are aware of themselves, and a lot of people will eat the hat.
point
The difference between a computer and a human is growing, as we move further and further away from logic and mathematics, areas that can be examined, compared and measured. And as much as thoughts are intertwined with hereditary factors, character, education, environmental influences, health status and more.
It is ridiculous to believe that we will develop a program that will think and act as a person, while we are very far from understanding how his thinking system works.
Why in the future? It already exists today:
http://www.youtube.com/watch?v=1QPiF4-iu6g
In the future, brains of mice, or something else, will be connected to computers and that's how artificial intelligence will be
I have proof that the computers on the science site have intelligence
First I thought about giga and I got mega, then I thought about Sabdarmish and I got some 4 in the middle. Is it possible that I was wrong or that the site's computers are trying to give me a message??
Food for thought
Sabdarmish Yehuda
Joseph
You're right. I thought giga and got mega.
(:))
Shabbat Shalom
Sabdar4mish Yehuda
@Yehuda Memory in computers is no longer measured in megabytes... Even on smart phones it is already half a gigabyte and on a personal computer with less than three gigabytes of RAM you cannot run Windows.
to someone-
Because feelings are not just relationships between inputs and outputs.
The best chess player in the world is unable to defeat new chess software running on a state-of-the-art home computer, let alone chess software running on a supercomputer a hundred thousand times faster.
In this sense, a computer is much more intelligent than a person in a chess game. Say what you will.
"It is clear that even though the computer knew how to deal with emotions, that is, it knew how to react to different situations similar to how a person would react as a result of their emotions, it is clear that it will not be able to really "feel", but only react according to the weighting of the data learned from the person and entered into it."
Why is it so "obvious" to you? If humans can feel emotions why can't computers? Huh, because we have a "soul" that computers don't? Because we have some magic powder inside our heads?
If they build learning computers based on neural networks that operate according to the same principles that our brain works on, I see no reason why they should not have both emotions and self-awareness, just like us.
PS - just out of curiosity, what does the name you chose to write under mean?
The article was published in 1987 when computers had a memory of several dozen K. I had the Commodore 64 which had 64K of memory, today we are talking about personal computers of several megabytes. It amazes me how Mazar Haim outdid himself even then and if it weren't for the date shown at the beginning, one would think it was a current article.
This problem is also a topical problem, for example, the winner in Israel of the short three-minute lecture competition brought in her lecture a classical work written by a computer and discussed the subject.
Well done stranger!
Good night
Sabdarmish Yehuda
I actually believe that somewhere in the future they will succeed in inventing a computer that can imitate the capabilities of the human brain. For this a tremendous progress is required in all sciences; Autonomy, psychology, social sciences, even philosophy, and of course computer science. For example, the example of expressing messages in the form of saying things; Surely at some point they will be able to determine exactly the array of sounds, their intensity and pitch, which are needed to express any type of message. However, it is clear that even though the computer knew how to deal with emotions, that is, it knew how to react to different situations similar to what a person would react as a result of their emotions, it is clear that it will not be able to truly "feel", but only react according to the weighting of the data learned from the person and entered into it.