The analogy from 0 to 1, and then from 1 to 100 is a good description of the development of technology and products. For example, Ford made the first T-car and thought it was the process of completing the car from 0 to 1. After that, 100 years until now can be considered as the perfection of the car from 1 to 100. Although the current car is far more precise than the version 100 years ago, it still has an engine, a transmission, a steering wheel and four wheels.

The shape of the 1919 Ford Model T, from which we can find almost all the key elements of a modern car, even its license plate is the same as the modern license plate.

Source: Wikipedia.

The computer has this process. As mentioned above, its theory is based on Turing. After Turing, von Neumann designed an architecture to make this thing. So far, the computer can be considered to have gone from 0 to 1 the process of. Since then, there have been computers ranging from tubes to transistors to computers using large-scale integrated circuits. The shape is also larger than that of a house. In this process, its computational performance has turned many times, but still requires the input and output system defined by von Neumann and the central processing unit.

If we compare artificial intelligence with the above two, we can find that artificial intelligence has not finished the process from 0 to 1. There is no clear theoretical basis and no clear definition of artificial intelligence. We don't even know what intelligence is. It is precisely because of this that we have reached a certain goal first, such as defeating humans in chess, and then it is considered that this is not a smart situation.

The inherent development of artificial intelligence has changed many times. For example, people initially thought that this kind of intelligence relies on a physical symbol system, which works by processing the data structure composed of symbols. Humans seem to be like this, but this is obviously only a phenomenon, and at this level does not produce true intelligence. Next, we also considered the genetic algorithm, which is to follow the process of biological evolution, let the program have some random changes, and use a selection process (the survival of the fittest of the creature) to maintain the seemingly useful changes, but limited to various "if how There are too many combinations of this, otherwise it did not produce any valuable results at the time.

Learn about it: from 0 to 1 on artificial intelligence

Then at least some people began to turn to the neural network. This direction appeared very early, appeared in the 1960s, but it was not mainstream until it was paid attention in the 1980s. Jeff Hawkings I think this is related to the failure of the mainstream mentality of artificial intelligence at that time. Neural network researchers try to generate intelligence in the connection, which is inspired by the brain, because the brain is made up of neurons, which is an obvious fact. Recently, the most popular deep learning is in this direction, but this direction has not yet completed the process from 0 to 1, but has achieved remarkable results in specific fields (such as speech recognition, image recognition). In the field of Da Niu Yan Le, there were several key restrictions on deep learning, including:

● Lack of theoretical basis. Without this, the deep learning method can only be regarded as a black box, and the confirmation of most conclusions is determined by experience rather than theory.

● Lack of reasoning skills. Deep learning techniques lack the means to express causality and lack the means of logical reasoning.

● Lack of short-term memory. Compared to deep learning, the human brain has amazing memory functions.

● Lack of implementation and unsupervised learning. Unsupervised learning dominates the learning of humans and animals. We can discover the inner structure of the world through observation, rather than being told the name of each objective thing.

- This is what was mentioned in the June 2015 speech, and InfoQ has a relatively complete report on this.

From this reality, it is really difficult for us to say that artificial intelligence has completed the process from 0 to 1, but can only say that it is still in a state of blindness, although the occasional touch is more useful than before, and at the same time experience The area that we have accumulated makes us feel bigger and bigger.

But unlike cars or computers, it is widely believed that the process of artificial intelligence from 1 to 100 will happen very quickly, and it will not develop like a car for 100 years, and the computer has developed for more than half a century. If we can really solve some basic problems and complete the progress from 0 to 1, the process of artificial intelligence from 1 to 100 may indeed be realized in a few years or even days. The Internet and smart hardware have prepared sufficient data materials and infrastructure for this.

If the general-purpose strong artificial intelligence of the humanoid is really realized, then the next process from 0 to 1 will be faced. This process is no longer a mere technical problem, but will challenge our ethical concepts more, and even what is the basic understanding of life.

If human beings are not hypocritical, then they must directly admit that the fundamental purpose of our creation of artificial intelligence is to obtain a loyal and loyal servant. In the middle, there may be some emotional needs, such as those shown in the film "Artificial Intelligence". Love. But the latter is secondary, just like many children like plush toys, but these plush toys can be abandoned without psychological burden. We don't want to create a kind of creature that is completely equal to us. In that case, it is enough to directly breed or clone. We don't want to create a kind of transcendental intelligence that is above human beings and rules humanity.

From this perspective, strong artificial intelligence can also be divided into two phases, to complete the following two mutations from 0 to 1:

The first stage is a stage with intelligence but no ontological awareness. This stage of artificial intelligence will fulfill its role as a loyal servant. This is the stage from 0 to 1 that is currently going but has not finished. The end of the process is that artificial intelligence becomes the vast servant of humanity and occasionally acts as a partner. We are currently at this stage.

The second stage is the stage of generating my own consciousness. At this time, it may begin to ask for its own existence value, thinking about why humans have so many emotions. At this time, people may still have an advantage in imagination and judgment. If we believe that thinking is inherently extroverted and seeking freedom, conflicts are likely to happen, because at this time you can't ask it to exist only within the scope you set, it has its own choice. It's completely unpredictable at the moment, and this process from 0 to 1 will start and when will it end.

Compared with the former, the new era that the latter is going to open is far-reaching and wide, and it does bring a lot of new problems.

These two processes from 0 to 1 imply the following three relationships between human and artificial intelligence:

(1) The ideal state is the master-servant relationship, and artificial intelligence serves humanity. The robot is a human robot. This is very simple, and it will definitely be the case in the first phase of artificial intelligence. We can also think of it as anthropocentrism, where artificial intelligence only pays but does not enjoy rights.

(2) The worse is that the robot generates consciousness, and there is a human-machine moral problem at this time. It is also a life, but this life is made by human beings. How should humans get along with it? Another troublesome branch at this time is that the human machine may need to be fused. In order to prolong life, get rid of disease, and travel in long space, many people may actively become half-human and half-machine. How do you define the rights and responsibilities of people, semi-robots, and robots at this time?

(3) The worst situation is to create super intelligence. At this time, this superman must be placed on human beings. People either turn into such superhumans or destroy them. If you can go to the second stage, it is easy to evolve into the third stage. The third stage can be considered as the result of the second stage of development from 1 to 100.

Americans like to do A. I. And I. A. The controversy, countless big scientists are involved, the former feels A. I. It is an independent intelligence, while the latter feels that intelligence on the computer is aimed at better serving humanity. But in fact, the difference between the two is not as valuable as it is. Before the artificial intelligence is really conscious, the two are in fact unified, and they need to improve their intelligence quickly so that they can better help human beings.

This issue will be more interesting from a philosophical point of view. The organic body of human beings is actually very fragile, perhaps to compensate for this fragility, so God has given humans a higher intelligence. Mr. Wang Dongyue believes that this process can be called weak compensation. If this is true, then the possibility of super intelligence finally does not exist, because this creature has both a strong body and super high intelligence. Nearly perfect living body in the universe.

It took mankind nearly 10,000 years to reconcile the contradictions between people. Only then did it realize that freedom and equality were better choices 200 years ago. Although it is still impossible to avoid killing and killing, modern civilization has achieved basics at many points. Consensus, the difference is only to the extent. If the latter state is reached, then the equality of man and machine needs to reach a consensus faster than this speed, otherwise the earth may not be able to withstand too much toss.

Stainless Steel Non-Magnetic Wire

Stainless Steel Non-Magnetic Wire,Cable Wire Railing Stainless Steel,309L Flux Core Wire,Marine Grade Non-Magnetic Stainless Steel Wire

ShenZhen Haofa Metal Precision Parts Technology Co., Ltd. , https://www.haofametals.com

Posted on