A tentative discussion on brain thinking and future computer vision (3)

Author: Zhang Shuai
Our brains are made up of 100 billion neurons, each of which has no intelligence, but only when so many neurons pass the transmitters at the speed of light (brain waves) and transmit information in their own language (encoded). This is a creature that has evolved over billions of of years, and that primates have the intelligence to transcend instinctive responses.
Human intelligence is based on neurons in the brain cells, but the neurons do not learn and train after the day, and the intelligence can not grow. Human intelligence is embodied in three aspects, one is the perception of time, the second is the introspection of self-development, and the third is the innovation of knowledge and skills.
And the future of the computer fully equipped with the conditions for the growth of artificial intelligence, I refer to artificial intelligence is not a weak artificial intelligence, but can be completely through Turing test of strong AI, even super AI. I think that once the computer has a strong AI, then it is very easy to have super AI, but from the weak AI leap to strong AI, is a hurdle, I estimate it will take 50-70 years from now, that is to see this article most people in the lifetime can be seen.
Why is the future of the computer fully equipped with artificial intelligence growth conditions.
First of all, the Internet allows computers (in this paper, computers not only refers to computers, as well as tablets, smartphones, in short, all intelligent terminals in this article collectively referred to as the computer) like neurons can exchange information, transfer "recursion." When the computer connected to the Earth has reached a certain scale, each computer's computing power beyond a certain scale, the information exchange of broadband over a certain scale, then the large-scale computer combination is like a neuron combination of the brain, from quantitative change, with artificial intelligence.
For this condition, we have the IPV6, which allows the number of interconnected computers to be more than the sand on Earth. At the same time, we are in the development of mobile Internet, so that the real life, the number of smart phones grow rapidly. Our fiber is being upgraded, the transmission bandwidth is generally up to MPs (megabytes per second), and the future can even reach GPs (gigabytes per second).
These three technological changes are the basic conditions for the future of computer artificial intelligence.
Second, the advent of the Big Data era allows computer AI in the next 50 years, there is a breakthrough in weak Ai to strong AI bottleneck capital. Big data is not equal to large amounts of data, big data is on a large amount of data, there is a qualitative improvement, if there is a mature knowledge learning algorithm, based on big data to train computer clusters. This process is no less than the process of human learning of knowledge in society.
Now to the future of computer AI, the most lack of anything.
I don't think there is a mature algorithm for training computer clusters.
Recall the three traits of human intelligence mentioned earlier, time perception, introspection, and knowledge innovation. Then we also need an algorithm to train the computer cluster has these three characteristics, that is, if the computer cluster has these three characteristics, then he can reflect introspection, knowledge innovation, awareness of the time to enhance their own intelligence, this through their own self-improvement of the process itself is the embodiment of strong artificial intelligence.
So what exactly is an algorithm like this?
A TPM-like module is required to record the various state and calculation results of every part of each computer in a computer cluster, and to make these results uniformly in a computer cluster, a comparison collision of different algorithms for the same problem, and a migration of different algorithms for different problems. These algorithms have the above-mentioned module record, can make the same computer storage of the same problem at different time period of the calculation process, results, status, can also be different computer storage of the same time the calculation process, results and status, even different computer storage of different time of the calculation process, results and status.
The essence of human introspection is a process of reviewing and re-evaluating the past behavior and thinking of human beings, by comparing and assessing whether the past behavior and thought are optimal, and then correcting the behavior and ideas in the future when encountering the same situation in an attempt to achieve better results.
As the above-mentioned module can record the computer history of the calculation process, at the same time, multiple computers run the same program of the calculation process to compare and collide with the calculation of the optimal calculation process to correct other computer programs.
There are two kinds of human knowledge innovation, one is not according to the conventional sub-process but the combination of generating a better result of the overall process, the second is the experience and summary of other scenarios "migration" to the overall process, but achieved better results.
As the above-mentioned module records a computer history of the calculation process, the application of his algorithm to another different calculation process, one may be the calculation may be completely wrong, the other may be computational efficiency and time is better, for the former, try other computational process "transplant", the latter, you can "transplant" To the results of this algorithm, a new algorithm is innovated.
When the computer artificial intelligence realizes the day, we recall 50 ago now, will discover, originally the mobile internet, the cloud computing, the big data, the artificial intelligence algorithm, the image speech recognition, the trustworthy computation all is in for its birth to make the foreshadowing, is the mysterious, or God's arrangement.
Maybe Einstein was right about that.
When people are immersed in the progress of science and technology, Einstein is worried about the future of mankind, science and technology not only bring happiness, but also bring disaster, because our EQ is far from catching up with IQ, our morals are far from the destruction ability.