Li Deyi: Tools have always been two-sided, good for science, artificial intelligence, and awe.
Many people are asking me what is artificial intelligence? Why does the State Council want to introduce the "New Generation Artificial Intelligence Development Plan"? I personally think that there are two levels here. The first is to talk about the impact of artificial intelligence as a new engine on its own industry. The second is to talk about the impact of AI as an accelerator on all walks of life.
The history of the development of science and technology is the history of the expansion of human ability to understand the world and transform the world. It is actually the history of the development of labor tools. Today, as a new stage in the social era, the era of intelligence has moved from a power tool to a tool for developing intelligence. Data, information, knowledge, value or intelligence have become so important that our demographic dividend is declining, but the smart dividend is coming, so we are building artificial intelligence together today.
I want to introduce two typical cases, how it solves the problem of human real life.
The first problem is AlphaGo. Of course, it is not a go player, because the person who grabs the piece is helping, we temporarily call it the Go brain. The speed of AlphaGo upgrade is impressive. Take Li Shishi as an example. He studied chess at the age of 9 and got a professional three-stage at the age of 16. He took a career nine-segment at the age of 26 and spent more than ten years. Look at Ke Jie, who is a 5-year-old chess player. He won three championships at the age of 18 and spent more than ten years. AlphaGo only achieved the victory over the human champion in less than two or three years. Is there anything worth pondering here? That is, with the expansion of the space of Go, the Go and the human Go players will adapt faster.
Archaeologists have discovered more than 4,000 years of Go history, from the 12&TImes;12 chessboard to today's 19&TImes;19 chessboard. For such a huge space, people suddenly found that AlphaGo defeated us.
What deserves our deep thought is that if we change the rules of Go, we will expand the board a bit and use 21 & TImes; 21 board. Programmers can change the parameters of the program without spending too much effort. AlphaGo can play chess with humans on a 21&TImes;21 board. For a person, the human brain is difficult to cope with.
A recent important event has taken place, and the new generation of Go AAI AlphaGoZero has defeated AlphaGo. AlphaGoZero's system can learn through self-game, which can be said to be self-taught, so the public is surprised.
If the programmer parasiticizes the AlphaGo program and the Watson program on a single machine, letting the machine compete with both the Go and Chess human champions, it will defeat the two human champions. And if we really want to cultivate a creature who is both a Go champion and a chess champion, how difficult is it?
The second typical case is driving the brain. I have been engaged in unmanned research for more than ten years and have a deep understanding of driving the brain. The human driver, after a rookie and maturity, eventually becomes an old driver and probably needs to drive 30,000 kilometers.
The upgrade speed of the driving brain version must be greater than or even greater than the evolution speed of the human driver's driving skills; the environmental adaptability of the group intelligent driving must be greater than or even greater than the driving ability of a single driver specific scene.
For the unmanned car I am doing now, it is very familiar with the traffic geography of Beijing. We call it the “live map†of the driver. But if one day I came to Hefei, just change the map in my driving head to Hefei, then I will soon be familiar with the whole Hefei.
For a social person, it is not a simple matter to put Hefei’s map in his head. Therefore, the brain of Go is good, and the brain is good. It is very important to give us inspiration. My conclusion is that tools have always been two-sided, good for science, artificial intelligence, and awe. I have been thinking recently that driving a brain is different from Go, and I want to talk about three things.
First, the equality, the two sides of the Go brain are very equal, whether it is the sunspot or the whites are the same;
Second, transparency, because the rules are clear, very reasonable and transparent;
Third, the outcome is very simple, so this is a deterministic formal advantage. For the driving brain, it is more difficult to ask for it. It is neither equal nor transparent, and it is more uncertain. Therefore, I personally think that it is much harder to build a driving brain than to be a Go.
The question I asked the most is, will artificial intelligence one day defeat humanity? This question reminds me of men and women in this world. They are the opposites. Do men have to defeat women? Does a woman have to beat a man? This is basically a pseudo-proposition. The same is true in the future. A group of people and their robots form a certain degree of opposition to another group of people and the robots they make, rather than putting people and machines together to form a confrontation.
Artificial intelligence has become a new entry point for the human cognitive world to transform the world and become the most important source of income for advanced society. Someone once said that artificial intelligence is strong and strong. If the study of brain science can improve the cognitive efficiency of people's learning, it will be a major contribution of human beings.
Li Yanhong has a point of view. He said that China's artificial intelligence is not leading, and it is really impossible to say. But foreigners don't necessarily see it this way. A professor in the United States mentioned that China has a population of 1/6 of the world, but it does not have 1/6 of the world's smart resources. Therefore, we must enhance the contribution of innovation-driven development sources and find productivity in artificial intelligence.
What are the current industries with the greatest impact of artificial intelligence? I think one of them must have manufacturing. Manufacturing is the backbone of the real economy, and the automobile industry is the backbone of manufacturing. There are 2.1 billion cars in the world, and if every car needs intelligent voice assistants, there will be a large market size.
When it comes to the education industry, is it "tactful" or not as good as people? Rote on the back, it must be that the machine is better than people. Many people still think that education is the accumulation of knowledge. For teachers, education is the transfer of knowledge. For students, education is the mastery of knowledge. The knowledge of stocks in the human brain is conducive to the development of curiosity and imagination, but it may also restrict the development of curiosity and imagination. In the era of information intelligence, we must develop students' ability to acquire knowledge, decision-making ability and innovation ability.
Now that the first grade has begun to join the general education, how should the college entrance examination change? I put forward a hypothesis for reference: artificial intelligence can trigger the first step of the college entrance examination revolution, that is, through the online candidates dialogue, find the colleges and professions that they are interested in, submit their personal learning situation, and submit the information processed by robots instead of people. .
If artificial intelligence can detect the ability to acquire knowledge, decision-making ability and innovation ability, then I think it can be further extended to all walks of life, including the examination of our civil servants. Flipping the classroom, MOOC, micro-class, personalized teaching will definitely have more coaches to replace teachers. So young teachers, you must also have a certain sense of crisis, otherwise your work will be replaced by robots. The education dividend is the biggest dividend in China's demographic dividend. It is the foundation of China's becoming an artificial intelligence power in 2030. It is not far from us in 2030.
The impact of robots on humans will far exceed the changes that computers and the Internet have made to the world over the past few decades. The machine was invented by us humans, but in the end the machine may be stronger than us, and the wisdom it will come from may be stronger than our individual or group of people. And we want to make our own robots, not to make us the largest market for American computers. By 2030, according to the development plan of the State Council, our artificial intelligence is to lead the world. At that time, we humans will be more dignified and live more elegantly and intelligently. We will usher in a new era of dancing with people and robots. With elegant dance music, there is no doubt that human beings are always the leading dancers.
Automotive Connector 27P External Thread Assembly
Automotive Connector 27P External Thread Assembly
Automotive Connector 27P External Thread Assembly
ShenZhen Antenk Electronics Co,Ltd , https://www.antenkconn.com