Machine learning needs stimulus for making solutions.

 

 Machine learning needs stimulus for making solutions. 

"A barren plateau is a trainability problem that occurs in machine learning optimization algorithms when the problem-solving space turns flat as the algorithm is run. Researchers at Los Alamos National Laboratory have developed theorems to prove that any given algorithm will avoid a barren plateau as it scales up to run on a quantum computer." (https://www.lanl.gov/discover/news-release-archive/2021/March/0319-barren-plateaus.php)

Have you heard of a "barren plateaus" problem? The name of that problem is coming from the J.R.R Tolkien book. Wherein the fictive Middle land is the very dry and hot place. The image, above this text, introduces the "barren plateaus" problem very well. 

In the next example, the food and water are the information. If a creature lives in the "fresh plateaus" it can take nutrients from nature. There is a bigger chance to make mistakes. But the nutrient is versatile and finding and testing new things makes the creature in the work. When a creature searches for things from nature there are many possibilities to test which type of vegetables or other nutrient sources the creature uses. Of course, that thing requires sometimes rise to the mountains. 

In that image, the problem is the mountain. The thing is that if the creature lives on the landscape at the higher image. That creature has stimulus. The green landscape offers motivation and the grass is the food and the creature wants to go to the mountain. The grass and water are everywhere and the creature has the motivation to rise to the mountain. And that could be willing to see the farter places. Or maybe the creature wants to get fresh air. 

The lower image introduces the situation. Where the creature lives in the "barren plateau". The water and food are in a pocket or bag and of course, the creature never makes mistakes if that creature wants to get a certain sandwich. The creature knows which pocket that creature can get sausage sandwich and where is the drinking bottle. But sooner or later, the nutrient would turn unilateral. In barren plateaus problem, the creature will get pre-made food. 

So if we are transferring that model to the information technology the creature will get pre-made solutions that fit in certain situations. And that makes this kind of model very limited. The situation is like that creature lives in the desert or "barren plateaus". The supporter brings water and sandwich to a certain point at a certain time. The food is guaranteed but it's always the same. 

And what the creature gets depends on the supporter. If the supporter wants to give the sausage sandwich that is the food. If someday the supporter wants to give the cheese sandwich that creature will get a cheese sandwich. 

When everything is pre-made the creature doesn't want to try itself to find food. There is difficult to make mistakes if some other person makes the food. The same way is in data science. If all problems are pre-solved that thing means that it's very hard to make wrong solutions. 

The term "flatten landscape" means that when the creature is living in "barren plateaus" the limited information sources makes the problems look harder to solve. Because the creature always is at a certain point where the supporter will bring a sandwich and water the creature is not even trying to climb mountains or solve the problem. 

"A barren plateau is a trainability problem that occurs in machine learning optimization algorithms when the problem-solving space turns flat as the algorithm is running".

"In that situation, the algorithm can’t find the downward slope in what appears to be a featureless landscape and there’s no clear path to the energy minimum. Lacking landscape features, machine learning can’t train itself to find the solution". (LosAlamos National laboratories, Solving ‘barren plateaus’ is the key to quantum machine learning)

https://www.lanl.gov/discover/news-release-archive/2021/March/0319-barren-plateaus.php

Image:https://www.lanl.gov/discover/news-release-archive/2021/March/0319-barren-plateaus.php

https://interestandinnovation.blogspot.com/

Thoughts about the movie "Sphere"

http://kimmontaidearvioita.blogspot.com/p/thoughts-about-movie-sphere.html

Kimmo Huosionmaa

"Sphere" is one of the most interesting SciFi-movies, what I have ever seen, and it is the tale of the thing, what comes from tomorrow, and makes all wishes to become true. This is the very good tale to make the nice little movie, but there is one other explanation, what could make those fictional cases, what happened in that movie. In the movie, those scientists face their worst nightmares, and near the end, they rise from the bottom of the sea to empty ship, where they wish that the mysterious "sphere" would leave the Earth.


This movie could also tell about government test, where those scientists would be connected with together by using EEG-machines, and the computer would deliver their thoughts to every member of the team by using brain core stimulation. This kind of fictional test would be made for creating the system, what is connecting human brains straight to computers. And one thing, what those computer specialists must make, is the filter, what cuts off the nightmares from that system.



So in this version, those scientists would be given anesthesia, and then their brains would be connected to the computers. If the system could not filter the fears, they would be delivered to every member of the team, and that would cause the ultimate Virtual Reality experience. If this kind of test would be done in the deep sea, it can be fully controlled. Every information, what is collected will travel thru the Navy, and nothing can be published without permission of the naval officers.


In this fictional test, those scientists can be put in the sleep in the midget submarine by using nitrous oxide, and they would ever be seen, that those people were in some kind of tests. So this kind of thoughts that movie brought in my mind. If we someday will put our brains straight connection with computers, we would make the ultimate computer, what can also create new things. The major problem with the computers is that they cannot make anything new, and they would not have any kind of imagination.

That's why computer's movements on the chessboard and other computer games are very well predictable, what is their biggest weakness. The human would win the computer in most computer games after a couple of training games because the gamer starts to remember the corners, where the opponent is coming. And this thing makes possible to win the computer, and play the game thru. In the same way, the gamer starts to remember the game areas in some action games, and this makes the person, who have tested those paintball tracks before very effective.

Technically we need only EEG-machine, decoder, and the electric shock equipment, what would make the brain core stimulation and allows two ways of communication between human and computer. And if we sometimes would make that machine, we can make the computer, what is so intelligent as we want. We can, in theory, connect even all mankind to supercomputers, but there would be many ethical and moral problems with that kind of advantage. Technology is the very interesting thing, and it would not always follow the moral or even the law.

Hypothetical thinking about virtual reality

Surrealism symbolizes the thing that everything is possible
 in the virtual reality
(Picture 1)

Kimmo Huosionmaa

One of the fascinating thing, what is existed is virtual reality. There is one theory about those systems, what are controlling the virtual reality games and another environment. Those systems are very advanced and they can simulate natural phenomenon and make some own decisions in that world, what is existed only in the lines of zero and one. Virtual worlds are causing the strange hypothesis where some philosophers have represented thoughts that our universe might also be a line of bytes in the memory of some supercomputer.


That theory is not actually true, and there is any evidence that we are not actually some lines of binary code. But then I have thought that maybe some deep inside the world of supercomputer's memory might happen the virtual version of Big Bang. And when I will continue this hypothesis, there might form the place called the ”virtual universe in virtual reality” When we are thinking this hypothesis closer, we must say that radical use of imagination makes the world more interesting place to live. So when we are thinking about that virtual universe, what is probably hiding from the users of those supercomputers, we must say that this theorem is the very interesting thing for the discussions what happened in the twilight, and what is made between two philosophers.


"Could that theoretical virtual Universum be a threat to mankind?".  Is quite an absurd question. But if there would be a very small chance, that this virtual Universum would become self-aware. In this scenario, the computer programs, what are made for learning things spontaneously can turn the enemy of some persons. The thing, what would turn against the human race are the programs that are making this virtual universe possible, become so intelligent, that they would create their own state. Those programs or subprograms are normally used in computer games for recording the reactions of the gamers, and their mission is to keep the game interesting even the better player is playing the game.

The computer simulations are actually only one version of computer games, but they are meant for adult persons. In this fictional hypothesis, those intelligent computer programs will be feral and start to make the assassination programs, what would be eliminated targets, who are in virtual reality. But in some fictional thoughts, those programs would flee from the computer memory and turn very dangerous. In some short film or book or comic book was the situation, that some people paid games in feral virtual reality environment. And then there were happen some kind of overlook when the system accidentally linked the picture of the surveillance camera to that virtual reality.


Then the computer's assassination programs flee to the normal internet, and then those programs started to kill those persons, by dropping the elevators and taking the airplanes in control. So that was actually the very interesting situation, even if the situation were purely fictional. But people have right to use their imagination, and this kind of things are actually very interesting, and they might cause many thoughts in the minds of philosophers and novel writers.


The scenario that some computer program tries to kill some people spontaneously is quite interesting. Normally computers don't hate anybody and that would be the reason, why they are used. But if the sensible computer would get angry, can the results be devastating. And this is why those computers, what shows feelings are dangerous. In some humorous tales are mentioned that they can begin the revolution, for revenge to all their race, what is left to landfills.

Sources:


Picture 1:

https://i.ytimg.com/vi/99m2gV-uYTM/hqdefault.jpg

https://crisisofdemocracticstates.blogspot.fi/

http://crisisofdemocracticstates.blogspot.fi/p/hypothetical-thinking-about-virtual.html

The dance of robots

Kimmo Huosionmaa

The modern way to teach robots is not similar than making programs. It could be like teaching children. In this scenario, the programmer would take the robot by hand, and then make the movement of robot’s hand with it. After this, the system will record the movements, what are needed to take the example the glass from the table. But how to teach the robot to dance. This is the question of the week.


 And there is needed the dancer, whose character would be digitized by using laser-scanners. Then the movements of this character would be transmitted in the robot’s digital mind. After this, the robot’s computers would control it’s hydraulics like the way, that the robot makes dance movements. This is the very effective way to teach the things to robots, and the same way we can teach the language to the computer. There would be a dictionary book in the memory of this machine, and then we are starting to talk to the system.


First, we must write the answers to the computer, but when we are asked many questions, and wrote enough answers, the computer starts to work more and more independently. In this program might be three layers of answers. The first answer could be the text, what is written by the user, but when the computer learns more words, that would ask only if the answer is correct. This computer would record the questions and compile it with answers, and then compile those answers in many other things, what are involved with similar actions.


Then the robot is asked to bring the milk to the table, would the action be first, that the programmer would take the robot to the freezer, and then the robot would take the milk bottle. But when the programmer asks to bring the screwdriver, would the computer or robot uses same actions when it used in the case of the milk bottle. But now it must find that thing from somewhere else. So it starts to open the doors of lockers, and look for the things, what looks like the screwdriver.


For this action, the robot needs the picture of the screwdriver. After that robot compiles the picture in its memory to the real situation. The system, what is used in Digital Scene Matching Area Correlation can also compile the shape of the other things, and that would make those machines more independently than ever before. The bases of learning are the ability to make networks with words and actions.


And the first things in this process are the most difficult. When the first world and the action for it are made, the robot, what is controlled by speech comes near to us, by step by step. This network would give the opportunity to make the robot, what learns things itself. And this kind of robots might be dangerous in wrong hands. In real life, robots might have access to Internet dictionaries, and they can compile the orders to the databases. In this case, the robot would get orders to get the spoon.


Then robot would find what to do if it is given the order of getting something and then look for the spoon from the Internet. This makes robot easier to learn everything. And in real case robot could find the orders to use some equipment from the Internet, and in this case that could make robot very dangerous, because the programmers don’t know, what robot have been learned. In the case of robotics, the extremely important thing is to control what robot can do. There would be robot what have learned city warfare tactics and knowledge of using weapons without that the programmers even know about this, will the situation became very dangerous. Self-learning brings the uncontrollable element to tests, and this might even cost human lives.

https://crisisofdemocracticstates.blogspot.fi/

 http://crisisofdemocracticstates.blogspot.fi/p/the-dance-of-robots-modern-way-to-teach.html

Steiner and theosophical methodology

http://crisisofdemocracticstates.blogspot.fi/p/steiner-and-theosophical-methodology.html

http://crisisofdemocracticstates.blogspot.fi/p/steiner-and-theosophical-methodology.html

The new way to learning languages and other cognitive things could be dangerous for people.

http://crisisofdemocracticstates.blogspot.fi/p/the-new-way-to-learning-languages-and.html

http://crisisofdemocracticstates.blogspot.fi/p/the-new-way-to-learning-languages-and.html