Saturday, November 10, 2012

the creep part 2


Hi friends,
I would like to start with Searle thought experiment – The Chinese Room Argument.
It raises some very important and fundamental questions which are inclined more towards metaphysics instead of artificial intelligence that’s why I thought that I should include and discuss it right this moment because of its relativity with the current topic.
Searle proposed an imaginary situation in which he is in closed room having two slots. One Chinese slips in a paper through one slot written a question on it in Chinese language. Searle does not understand Chinese himself but he has a library inside the room containing all types of the books about Chinese characters. He matches the Chinese symbols written on the paper with the books available in the room, finds the answer, writes it in Chinese by copying same as appearing in the books and slides it through the second slit. The Chinese man reads the answer and founding it perfect, he (Chinese) assumes that the person, inside the room, is well versed with Chinese though Searle himself does not understand even a single letter of Chinese.
Searle argued against strong AI saying that mere a successful programming of any system is not sufficient enough to call it intelligent even if that system is able show some results similar to human mind. Although he accepts the state of weak AI to some extent but very rigid in his stance that mental phenomena could not produced by computer programming. He further stated that brain causes minds and minds cause consciousnesses. There are endless arguments both in favor and against the Searle. It is so pervasive that even a book is not enough to cover it all so the single line abstract of the Searle’s argument is that machines do not think and understand no matter how advanced and sophisticated we made them and I agree with his conclusions in the broad sense though my reasoning is slight different.
In my opinion the Chinese room argument is much bigger issue than it appears on the surface. This thought experiment asks what the definitions of knowledge, brain, mind and consciousness are and how all these work. We have to understand two things before analyzing Searle argument; KNOWLEDGE AND INFORMATION. I feel that sometimes we tend to overlap both of them.
This phenomenon is very important and needs to be understood very carefully. Let me take an example. Every mathematician is familiar with Pythagoras theorem of a2 + b2= c2 and he uses it to solve problems. He can carry on successfully from this theorem further inventing many new theorems but his knowledge is incomplete just because of a simple reason that he has borrowed Pythagoras theorem and many other established formulas for his work without going through the process of inventing those so the effort and pain taken by his predecessors is missing from his experience.
I want to make this issue clearer by using very common and easy examples. There are many cases in our daily life in which we experience knowledge even without noticing it. Let us take colors for instance. If we ask a five year child to explain the blue color, what will be his answer? We all know that he has the knowledge and understanding of blue color but he would not be able to explain it. The explanation of blue color is equally difficult for adults just because it needs to be seen or experienced. There is no other way of knowing it. Furthermore, it is one of those types of knowledge that do not have any counterpart in our world so one would not be able to explain it to anyone else through an example. This is what I would like to call knowledge. It is not transferable because words cannot describe it.
A blind man by birth cannot understand how blue color looks even if the wisest person in the world spends his whole life explaining because due to the lack of visibility, the blind man will only get information not knowledge. There are numerous examples of this phenomenon in our daily life. Without experiencing the whole process one can get only information and we generally misunderstand information with knowledge in day to day life.
Knowledge is complete phenomenon while information is only an abbreviation of end result.
The great Einstein formulated general theory of relativity first but he was not able to explain gravitation at that moment. Sometime later he tried to cope it with the concept of cosmological constant but could not succeed. Then, after 15 years since general relativity, he became able to come up with the solution in the form of curved spacetime and general relativity. In that span of 15 years, Einstein must have thought of numerous solutions, would have examined them from different angles and then negated them for one reason or the other. At last he gave the right version. Now, one can think that he or she knows relativity but it is not true as we have not gone through the process. We do not know those endless ideas those used to pop up in the mind of that great scientist. We do not know on which grounds he rejected all of them. We are only familiar with the appropriate version but he knew all the unfit versions also besides the right one so his understanding and knowledge had covered much more space in comparison of a person who knows the end result only.
A truth, even if it is truth, is never complete for the sake of knowledge, until and unless one does not know what is not truth.
Let us assume that we know that a work could be done or a problem could be solved in a certain way and we can say that is a truth or fact. So, it could be said for sure that for each and every way other than the right one is not the appropriate way and this saying is also truth or fact. Hence, we must understand that there are two types of facts; positive facts and negative facts. The number of negative facts will always much larger than the positive facts as we all know that there could be endless ways of doing a work wrongly while the proper way is only one but one could not know the all wrong ways unless and until he tries for the right way. Knowing the wrong ways is equally important otherwise we miss almost all other facts and left with only one which is the right version. So, we see that knowledge is not just a single truth but it comprises of endless facts and those could be acquired only if we go through the process.
Hence, the process is more important than the result even if it (process) is not fruitful.
I also want to put it mathematically because we often tend to believe more in numbers than the text in today’s scientific world. Einstein took 15 years to reach the end version. 15 years means 5475 days or 782 weeks or 180 months. Given the versatility of that genius, we can safely assume that he got at least one idea in a month if not more and negated it in his mind. It means that he would have thought of 179 more ways of postulating the theory. In other words, we can say that he knew 179 wrong and one right version. Even if we take the most conservative approach by picking only one idea in a year, his knowledge stands at 15 times more than the others. So there is absolutely no comparison between the information and the knowledge.
Information is just a drop of the ocean of knowledge.
Now let us talk about the machines. In my opinion, there is no fundamental difference between super sophisticated computer which can beat world chess champion and a simple calculator. The basic problem with the computers is that they are open only to the information. There is no way of enriching them with the process of learning. The entity which is required to go through the process of learning is intellect or mind and without it acquiring knowledge is not possible. The functions and outputs of any other system other that mind are limited to the extent of the information supplied to it. All physical informative systems work on a very simple give and take rule. They just return or show what we feed them; not less not more. Although sometimes it appears that they are able to think, like in the case of world chess champion beating computer, but it is not true.
If we look carefully then we find that in this case, the world champion is not playing against a single computer, but all the chess playing minds those programmed the software consuming years. Actually, in objective view, it is clear that the super computer is doing nothing besides playing a role of a middle man between the world champion and the programmers. The actual game is being played between the champion and those programmers. The knowledge of the programmers is reflecting in the computer in the form of information. Computer may win the game but it does not proof that it (computer) has more thinking power or knowledge than the champion but win will be due to the reason that the total information fed by the programmers may exceeds the knowledge of the champion.
Sometimes, when information fed to any system becomes extraordinarily huge and pervasive, it (system) gives the impression to be intellectual. All informative systems, super computers which can talk and ultratech robots, even those which are in the wraps of the future, no matter how advanced and sophisticated those would be, will be informative only. I do not have any doubt in my mind whatsoever that we could not make those to think simply because we would never be able to provide intellect as it is not a physical thing and we cannot create it.
There is a lot of talk about the testing of thinking power of the systems; especially Turing test. This test claims that if any machine is able to answer questions in such way that resembles human behavior then it could be said that the machine is thinking. This test also sets a standard that if a human fails to identify the machine by reading its answers and mistaken it with human mind, then it is certain that the machine is thinking. I do not see any merit in this argument. As I said earlier, that when information of any system becomes enormous in comparison with human mind, it gives fake impression of thinking power. Given the development in computers and robotics in modern science, it is quite possible that many of them could be able to pass this test but, in my opinion, all this does not proof anything. We have to understand the meaning of “thinking” first before setting the parameters for the test otherwise the result could be misleading.
Thinking is a mental process. It takes a start from evaluation and analysis of anything. By doing this, thinking explores both types of facts regarding the issue; positive and negative. At first, it looks carefully at the first hand information of the subject, and then roams around the periphery of the core point. In the process, it collides with the other facts; it sometimes accepts them, sometimes amends them, sometimes negates them and sometimes adds to the tally by finding new ones. There are two basic characters of thinking; evaluation and evolution. Any system, which claims that it can think or has intellect, should be able to meet these standards. A thinking entity should have the capacity of testing, modifying and adding to the information provided to it without any help from outside the system.
Now let us judge machines. Machines cannot test the information. They accept it verbatim. There is no challenge from their side. If we replace all intelligent moves with silly and wrong move in the programming of chess playing super computer; it will not refute it; no matter how advanced chips it could have but it will play accordingly to the programming when the situation will arise. On the contrary, even an average chess player will not accept silly moves if anyone asks. Yes, computer can compare between two sets of information very precisely but it cannot decide by itself which set of information is right or wrong. If we tell it that the wrong version is right, it will act accordingly and vice verse. A thinking and intellectual entity will never do it. So there is clear-cut difference between thinking systems and informative systems. Machines do not fulfill even the other two conditions as neither they can amend the information nor add anything new to it by themselves. So, I am sorry to say but no physical system other than mind is able to cross the bar.
To summarize the issue we can say that information is not an alternate of knowledge simply because in the absence of intellect, no entity in the universe, can absorb any kind of knowledge. So, there could be no such phenomenon in the world that could be called artificial intelligence; neither strong nor even weak.
with love,
Howdoo and welcome back for another cruise on The Bad Ship Clusterfucknation! ;o)
Couple things:

No comments:

Post a Comment