maruxz (maruxz) wrote,
maruxz
maruxz

Is human memory 'better' than computer memory because it is imperfect?

The answer is “Of course!”. Why? because still I am a human and still consider human race the most superior on the Earth. Computer is the child of continuing human mind development from pile-of-lamps type apparatus to a very sophisticated mini-chip 'thinking' creature. We have to deal with the mind to describe the goodness or badness of the memories, because they are the part of mind which construct the identity.</font></p>

Let's find out why computer is worse than human in general, why people still consider computer as tool. The worst nightmare of modern computer engineers-Frankensteins is the future where computer based organisms can probably take over the humans. All these futuristic movies, sci-fi books dealing with computer mind incarnation to humans, animal-humans and machines or robots, has the feeling that the line of human superiority soon will be attacked by the creations of the same man. There we go: new mechanised race is going to take advantage of their intellectual and physical superiority and destroy lazy, mutant and pervert human race. We are afraid of our technological kids, we are afraid of emerging new mechanical interference to human body: cloning. Three years ago the chess match between the World's best grand master Gary Kasparov and IBM supercomputer Deep Blue showed humans that their fears are 101% real because human is beginning to lose to his creation. The throne of human goodness is not solid any more and the descendants of Mr. Alan Turing every time doing his test is scared, because soon computer mind will not fail to pass it. That's why we still think that our minds are better than computer ones, including the memories, because our memories are real and theirs - only combinations of 0 and 1.


MEMORY

Memory is the storing and retrieving information in human brain (Encarta Encyclopaedia). Memory is when human experiences are retrieved in the mind (Encyclopaedia Britannica). Even though we have this simple definition, the memory as an object itself is not clear at all, because from the beginning of thinking people were trying to find out why we remembering things we saw, heard, felt even dreamed. The process itself is crucial for thinking, learning, problem recognition and solving it, decision making and constructing identity.

The problem is (or is not) that memory of human by nature is imperfect. If one see a picture of a tree and after some time try to remember how the tree looks like, there cannot be the exact image of it produced by the brain. Some information is missing, but instead of missing part there are always other information taken from experience. Memory is a very dynamic process, where memories are reproduced by using experiences and other memories of other things, which also are not the same as it was in reality. How the experience is real (we know, that psychoanalytic can "implant" some experiences one cave never seen, felt or hear without a close contact with reality) is another question.


CAN COMPUTER THINK?

Unfortunately NOT. Still not, because today the super computers just store information and make just a little in order to combine data to make a sentence, to move robot's hand, or give the answer during the Turing test. As Charles Platt was speaking about robots which contain computer brain: "Today’s best robots can think at insect level" (Platt,1995) there are almost no hope that tomorrow the Turing test (The test is following: there are three competitors put in separate rooms and communicating by typing. One of them gives the questions to another two from which only one is human. Another one is supercomputer able to give answers to the questions, that means able to recognise the language. The purpose of questioner is to find out are both of replyers humans and which of them is the machine. Until now the test failed all the times) would be succeeded. Meanwhile the future promise a lot.

What the programmers already achieved and computers may suggest concerning specific language, memory and understanding matters:

#1. Voice recognition In early 1980's two software companies Dragon Systems and Kurzweil Applied Intelligence from Boston began selling the programs which enables PCs recognise their masters voice. In the beginning there were a few commands but until 1997 these programs could recognise about 60.000 words (Garfinkel, 1997). Three years of intensive work and research passed already and the newest product of this invention is... Erickson G28 mobile phone, which can be controlled using nothing but the voice. The problem of this achievement is that computer cannot understand the flow of words, so the pause after every word is still needed.

#.2 Digital speaking In 1996 Digital Equipment Corporation introduced a speech synthesiser for blind people named DECtalk, so the text can be read by computer. But "Today's computers speak pretty well when they operate within narrow parameters of phrases, but sound mechanical when faced with unrestricted English text. Real breakthroughs will require a better understanding of natural language. Give it five years" (Garfinkel, 1997)

#3. Image recognition MIT (Massachusetts Institute of Technology) AI laboratory scientist Tomaso Poggio has developed the program which can recognise a particular person from several group pictures and another which can recognise objects and faces from the line drawings. (Garfinkel, 1997) Still these programs are to be improved because they cannot deal with every pictures and objects, only very simple ones.

#4. Chess playing RS/6000 SP-based computer system which was designed for cleaning up toxic waste sites, forecasting the weather, modelling financial data, designing cars, developing innovative drug therapies and better known as the Deep Blue in 1997 beat the world's best grandmaster in chess Gari Kasparov. What was the reason? David G. Stork, the editor of book HAL's Legacy: 2001's computer as dream and reality (MIT Press)(http://mitpress.mit.edu/e-books/Hal/) explains: "Humans are slow but exquisitely good at pattern recognition and strategy; computers, on the other hand, are extremely fast and have superb memories but are annoyingly poor at pattern recognition and complex strategy. Kasparov can make roughly two moves per second; Deep Blue has special-purpose hardware that enables it to calculate nearly a quarter of a billion chess positions per second." (Stork, 1997) This clearly shows that computer has perfect memory and speed of operating the data, but not making strategies, even though Deep Blue had used the strategies: the positions of movements to win the game, gambits etc., but, of course, these strategies are the intensive work of dozens of programmers and not own.

It is very clear that computers still have to learn more in order to begin thinking and operating memories, not only information coded in 1's and 0's.


PROBLEM RECOGNITION AND SOLVING IT

The recognising the problem is the main principle humans deal with every day in relationship among people as well as people and nature (I mean when you are alone in a fields and suddenly the storm begin, you understand what is going to happen). The other thing, which requires even more intellectual capabilities is not only grasping the problem, but also solving it with least disadvantage to one's well being and existence (I mean, when you are running to the shelter when the storm is to begin). The recognition and solving the problems can be only if you have some memories from the past and experiences how to handle with possible and real problems. The question occurs if computer can recognise and solve the problem using their memories? They can check your spelling when you write a paper in Word 97 but because of installed vocabulary from which the program is taking similar words to wrong one. This is a very simple example, but what is happening with more complicated information? Do not forget that Deep Blue beside playing chess is also designed for cleaning up toxic waste sites, forecasting the weather, modelling financial data, designing cars, developing innovative drug therapies but not still for solving personal problems. When it can say how for this person is better survive the depression after mother's death, it will not be IT anymore. It will be something between machine and human while now the distinctions still is quite clear.


DECISION MAKING

To make a decision one need to predict the consequences of this decision to evaluate if it is right or wrong. Predicting how one or another process is going to develop or is going to be developed means creating strategies. Here memories are involved from another side because if one creates a strategy, there is a very important need to compare new one with previous, already used strategies; to see if previous strategies were successful or not and why they were such. See the history of the war. Alexander the Great conquered the Sparta and saw their war strategy called falanga, when soldiers march in square shaped troops and when the first lines hold long spears which do not allow enemy to get close very quickly. This strategy was successful, so Alexander went even to India. In chess game the particular set of movements are called after people or other names to identify different strategies they used or it has specific characteristics. Grandmaster can foresee half of dozen moves ahead while computer is checking possible moves in speed of 200.000 positions per second.


CONSTRUCTING IDENTITY

the question of what is identity, how humans get their personal identities etc. hasn’t been answered yet even though the discussion “who am I” began from the moment humans started to speak and think. Now we can say more about identity than our predecessors thousands of years ago. We know that memories are the crucial element of an identity. Without memories we cannot use the data and surrounding information; without memories we cannot make solutions to constantly arising dilemmas and problems; thinking strictly depends of memories; without memory we cannot learn language and express ourselves while expression is the main way to represent us as persons with different identities in modern society. I will not discuss the eternal topic if animal also has identity or not, because in the future another discussion will be crucial for survival of our society and even race: if computers and robots have identity and if do, is this identity superior to human one? Hans Moravec, cited above in Charles Platt’s article, predicts exactly such future when robots will obtain the identity witch will be as human, so will be able to replicate and take over the human race.

The official science still do not have much evidence that this scenario is going to come true in next 50 years, meanwhile science fiction writers and cinema created the futuristic image of cyborgs and smart robots.


CONCLUSIONS

Maybe it is still early to speak in official scientific way about humans replaced by their creations - computer-minded robots or cyborgs, because as we can see not very much was achieved towards constructing artificial identity using memories. The reason is very simple - lack of knowledge about human brain. It is so complicated, that we cannot understand where and how all the data are coded in the brain, what kind of measure should be introduced of storing information and transmission? When we will know how our brain function, we will be able to create something like artificial brain. Then cyborg era will be a reality.

All the time we are speaking in abstract terms, some of them even do not have agreed definition, about unreal things, which we are going to have in the future. It is not stable and our life expectancy is dealing almost with dreams and science fiction. I would like to share with new social nightmare which is already on the way. Couple of months ago British scientists finally received the permission to clone human for scientific purposes. Mouse, sheep and monkey has already be cloned and carefully examined. After some time when sheep Dolly was cloned scientist noticed that one of her DNA’s crosses is a little bit shorter than natural one so it influences the life longing. Of course, scientists promised to solve this problem, but who can guarantee, that cloned human will not be so perfectly copied as expected? Let’s say the human is cloned, he/she is perfect both in body and intelligence. A new race is going to appear then: cloned ones. The humans will be divided into natural ones: ugly, lazy, stupid, and cloned ones: perfect. Then will be the same scenario as if we had cyborgs in Hollywood movie or sci-fi book.

Of course, the question if cloned person will have mind, intelligence, memories and this thing people usually call soul, still remains unanswered and crucial for further development of scientifically improved person.


REFERENCES

Encarta Encyclopaedia, article on Memory (mental process) at http://encarta.msn.com/find/print.asp?&pg=8&ti=761578303&sc=0&pt=1

Encyclopaedia Britannica, article on Memory at http://www.britannica.com/bcom/eb/article/printable/7/0,5722,109427,00.html

Platt Charles, "Superhumanism" from WIRED No. 3.10 1995 at http://www.wired.com/wired/archive/3.10/moravec.html

Garfinkel Simson, Happy Birthday, Hal from WIRED No. 5.01, 1997 at http://www.wired.com/wired/archive/5.01/ffhal.html?pg=3&topic=

Stork David G., The end of an era, the beginning of another? HAL, Deep Blue and Kasparov article from IBM research page at http://www.research.ibm.com/deepblue/learn/html/e.8.1.html

Encyclopaedia Britannica, article on Learning at http://www.britannica.com/bcom/eb/article/0/0,5716,48642+1+47530,00.html

Subscribe
Comments for this post were disabled by the author