Shulou(Shulou.com)06/03 Report--

It has been more than seven years since artificial intelligence broke out again in 2012, people have gradually gone through confusion, panic and excitement, especially the Chinese public, because they have not experienced the two most exciting parts and falls of artificial intelligence in the 1970s and 1990s. it's particularly obvious. After years of precipitation, people have a deeper understanding of artificial intelligence under the guidance of deep learning.

Huang Tiejun, a professor at Peking University, said that after two ups and downs in artificial intelligence, it has been over a hurdle this time, and we human beings will enter the era of intelligence, not the cold winter of intelligence.

Today (November 3), Academician Pan Yunhe gave a report entitled "artificial Intelligence 2.0 and Digital economy" at the first open course of the Science Lecture Hall of Renmin University and the Hillhouse School of artificial Intelligence. In the report, Academician Pan Yunhe also made profound insights into the development of artificial intelligence.

Academician Pan Yunhe believes that since 2015, due to 1) the outbreak of new social demand, 2) great changes in the information environment, 3) great changes in the basis and goals of AI, artificial intelligence has moved towards the 2.0 stage. Based on this understanding, the state issued the New Generation artificial Intelligence Development Plan in 2017, which puts forward five major technological directions for the development of artificial intelligence: big data intelligence, swarm intelligence, cross-media intelligence, man-machine hybrid enhanced intelligence and autonomous intelligence system.

Academician Pan Yunhe then analyzed the essential reason why artificial intelligence is moving towards 2.0 from a philosophical point of view: we have gradually entered the "physical space"-"human society"-"information space" from the traditional dual space based on "physical space" and "human society". " The emergence of "information space" is bound to have an impact on all aspects of human survival, for example, through information space, human beings can transform physical space in new ways; through information space, human beings can know ourselves in new ways.

This kind of transformation is revolutionary for the whole human history. We are fortunate to live in such a historical stage, and we should seize the historical opportunity.

The following is part of the speech delivered by Academician Pan Yunhe, which has not been confirmed by me.

Note: Academician Pan Yunhe's report is divided into two parts. The first part talks about the current situation of digital economy and China's digital economy. The following is a part of the second part, "artificial Intelligence moving towards 2.0". Omitting the "artificial intelligence 2.0 technology has emerged", "the application of new generation artificial intelligence technology" and other contents.

63 years ago, in 1956, the concept of "artificial intelligence" was born in Dartmouth, USA. At that time, there were a group of famous scholars, including Professor McCrarthy of Stanford, Professor Minsky of MIT, Professor Simont and Newell of Carnegie Mellon (all four of whom later became Turing Award winners), as well as Shannon of Bell Labs and Rochester of IBM. At that time, they established the concept of "artificial intelligence": to enable machines to recognize, think and learn like human beings, that is, to use computers to simulate human intelligence.

Then in the 1970s, the field of artificial intelligence research mainly aimed at seven directions, which have been playing a very important role up to now. Include

1. Proof of machine theorem. Now it's logic and reasoning.

2. Machine translation. Now it's natural language processing.

3. Expert system. Now it's problem solving and knowledge representation.

4. Game. Now it's a tree search.

5. Pattern recognition. Now it's multimedia cognition.

6. Study. Now it's a neural network.

7. Robot and intelligent control. Now it's perception and control.

The fact of these seven directions is to simulate different human behaviors, in which many schools are introduced, the most typical of which are symbolic school, connection school and behavioral school.

Artificial Intelligence is moving towards 2.0 (2015)

Artificial intelligence has experienced a lot of twists and turns, time constraints will not elaborate.

In 2015, artificial intelligence began to rise rapidly from the low ebb. In 2015, the Chinese Academy of Engineering has a topic called "Research on the Development Strategy of artificial Intelligence 2.0 in China", in which we put forward an important concept, that is, artificial intelligence is moving towards 2.0.

What is the reason? We talked about three reasons.

First, the needs of society are different. In the past, artificial intelligence used to simulate a person's intelligent behavior with computers, which was done in the previous 60 years, but now there are new needs in the society. what we need to solve is the problems of intelligent city, intelligent medical treatment, intelligent transportation, intelligent games, intelligent manufacturing and so on. Most of these problems are not to simulate the intelligence of a person, but to simulate the intelligence of a group of people. What needs to be studied is the intelligent operation of complex giant systems. Therefore, human intelligence is very important, but the intelligence of the system is more important. For example, the transportation system, in the past, self-driving in intelligent transportation can solve the problem, but recently it has been found to be no good. In the end, the problem to be solved is the vehicle-road coordination system.

Second, the information environment has changed greatly. Sixty years ago, artificial intelligence aimed at the intelligent simulation of a computer, but now we are faced with the information environment of the Internet, mobile computing, supercomputing, wearable devices, the Internet of things, cloud computing, online communities, the World wide Web, search engines, these tools constitute a complex information giant system. If our artificial intelligence does not make full use of such a new information environment, it must be a weak artificial intelligence. our artificial intelligence must study artificial intelligence based on a giant system of computers, computer-connected networks, computer-generated big data, various wearable devices and mobile devices.

Third, the foundation and goal of artificial intelligence have changed greatly. Sixty years ago, artificial intelligence wanted computers to be as smart as people. Sixty years later, we understand that computers can be smarter than people in some ways, it can remember more than people, it can count faster than people, and we even find that it can play chess better than people. This shows that computers will certainly do better than people in some areas, just as cars must walk faster than people on the road, but people's legs play an irreplaceable role. Therefore, it is difficult for computers to describe some intelligence of human beings. If it cannot be described, it cannot be programmed, and if it cannot be programmed, it cannot be realized by computer. Now more and more artificial intelligence scientists realize that machine intelligence and human intelligence are two different kinds of intelligence, one is the intelligence of silicon chip, the other is the intelligence of human brain, the two must have their own advantages, and the best way is to combine them and use them freely, so that they can form an agent that is smarter than people and machines to serve people, which should be the new goal of the development of artificial intelligence.

Based on the above understanding, we believe that artificial intelligence must move towards a new generation. After more than a year, going up and down repeatedly, China finally released the "New Generation artificial Intelligence Development Plan" in 2017. This "plan" draws the world's attention to the Chinese people's determination and confidence in the development of artificial intelligence.

Another concern for the world is that the Chinese have put forward five technological directions for the development of artificial intelligence: big data intelligence, swarm intelligence, cross-media intelligence, man-machine hybrid enhanced intelligence and autonomous intelligence systems. In these five aspects, except that "swarm intelligence" is a concept put forward by foreigners, the other four are new concepts put forward by China. For example, big data Intelligence, foreign efforts are engaged in big data technology and deep neural network. We think that deep neural network is a kind of big data intelligence, and big data intelligence certainly includes others (such as knowledge graph). Big data and deep learning are not all of big data's intelligence. The same is true of swarm intelligence, we believe that in such an era, the wisdom of many people together to form a new intelligent system will be an important research direction. The same is true of cross-media, for its application, we propose to study intelligent city, intelligent medical treatment, intelligent manufacturing, these fields are to study the intelligence of complex giant systems. The state believes that it is not enough to study these, and that there is an intelligent economy. Renmin University has an important responsibility in this respect. In the future, economics and engineering will be combined to form a unified study of "intelligent economy."

The essential reason why AI moves towards 2.01.The transformation from binary space to ternary space.

After we posted this report, we conducted an in-depth study later, and found that in addition to the three reasons mentioned above, there was another fundamental reason: the previous world was binary space, one element was human social space (Human society space), and one element was physical world (Physical space); but in the past 30 years, the world is forming a new space-information space (cyberspace). Our world is changing from the original PH binary space to the composition of a CPH triple space.

In the past, we also had a lot of information, most of which came from human society. We wrote a lot of books, held many meetings, had media, television, newspapers, computers, and even used computers to process data. All these messages are sent by people at first, so it does not become a separate space, it is attached to human society. And then we connected this information, with the Internet, with mobile communications, with search tools, our information began to bypass humans, and we sent a lot of satellites to the sky, 24 hours a day. 3600 seconds an hour keeps sending messages to us; we have a lot of cameras on the street, and these cameras on the street are constantly sending messages to us. It is said that there are 300000 cameras on the streets of Hangzhou. 300000 cameras keep sending messages, people can't handle it, computers can't handle it, and it's hard for us to use it completely without turning these data into knowledge. This has formed a new era, which we call the big data era.

According to foreign books, the main feature of the big data era is that the data is very large, which is right. But I think not only the data is very large, but also the hidden characteristics, most of these data are not sent by people on their own initiative, they are sent out by the physical world, bypassing human beings. For example, your phone is sending messages all the time, and collecting this data can pinpoint where you have been. These messages are about people, but they are not from people, they are from things. This is the essential feature of the big data era.

These things prove that a new space has been born, which is the space of information, which is not attached to human beings, and is its independent space. In such an era, if data does not become great knowledge, we will not be able to use it; only by turning it into great knowledge with artificial intelligence can we control this era.

2. new changes in the formation of information flow after the change of information space.

In the original binary space (PH space), we have physical space and human social space. Human beings constantly study physical space: what is matter, what is the law of its change, what new matter it will become after change, and so on. All this knowledge makes up our physics, chemistry, biology and so on, which we call natural science. Natural science is the knowledge of physical space.

Then human society has to transform the physical space, how to design, how to process, how to maintain. The sum of all this knowledge, including architecture, mechanical manufacturing, electronic engineering and so on, we call it engineering technology.

Then human society has to study ourselves, our economic behavior, political behavior, cultural behavior, ethical behavior, all of which we call social science.

Now there is a new space, that is, the information space. Human beings have found that in addition to directly observing the physical space, the new side of the physical space can be seen through the information space, which is a very important change. For example, in the past, when we looked at the physical space, we could see how the traffic at the crossroads was congested, but in the information space, we could see how the traffic congestion at one intersection affected the traffic changes at the four surrounding intersections, which in turn affected the traffic changes at the sixteen intersections around us. In the past, we could not see it, but now it is easy to see through the information space. How does the development of a city affect the development of several surrounding cities; how does the development of a city affect the development of several small towns around it? in the past, these things had to be clarified by social investigation. Now it is demarcated by big data in the information space, which is transformed into the information space, and then the physical space can be directly transformed through the information space.

We can transform the physical space through the information space, the earliest example is the drone. If drones appeared ten years ago, I think the fifth-generation aircraft would not be what they are now; if there are sixth-generation aircraft now, they must be aircraft with a fleet of drones. A plane with a bunch of drones would be awesome. This is definitely a new mode of operation. In the future, it may be a person fighting a war with a batch of weapons, and these weapons will advance automatically, and people will become very important, and everyone will be an officer. Therefore, the way of transforming physical space through information space is bound to emerge in large numbers in the next decade.

We also find that through the information space, human society can re-recognize human beings themselves. China conducted a census five or six years ago, and I was told that the census took half a year. It took tens of millions of staff to determine dozens of parameters such as age, sex, place of origin, work, education, personnel relations, and place of residence of China's 1.4 billion people. This is a very huge project. But I believe that in five years' time, if we do another census, it will only take a few weeks, a few hundred people will probably be enough. We can use big data to put all kinds of data together, not only can we figure out the above data, but also what stage you are in and what things you like, all add to the population data. In another fifty years, it may be possible to paint a portrait of everyone.

3. New information flows will lead to new changes in cognition.

In the research of social science, a new realm will appear in the study of people, various behaviors of people and the relationship between people.

Therefore, the change of space (the change from PH to CPH) not only affects the discipline of artificial intelligence, this new change of information flow will also affect the new change of cognition, affect all science, engineering and social science, and provide new research methods, new research picture and new research channel for all social and engineering science. Some scientists say that science will have a fourth model. In the past three models, one is the method of observation, the other is the method of scientific calculation, the other is the method of computer simulation, and the fourth model is the mode of research through big data.

In addition, there will be many new disciplines, that is, complex giant systems. For example, urbanology, in the past, there was only one science that studied cities-- urban planning. Urban planning is a science that studies urban physical systems, and people in urban planning do not recognize it as a science, but admit that it is empirical integration, not science, and there is no theory. This theory is fragmented, not a whole theory. Therefore, we need to accumulate experience in planning, ranging from urban planning to community planning. Now Academician Wu Zhiqiang of Tongji University puts forward "to reshape the theory of urban planning all over the world", which is based on the city big data. Tongji University claims to have established the most complete city in the world, big data. If Renmin University wants to revive its economy, it should build the most complete economy in the world, big data.

This is the case with the urban operation system, and so is the environmental ecosystem and the health care system. Health, not only with hospitals, patients, but also health insurance, pharmaceuticals, home health, wearable health equipment, pharmaceuticals, is the human health system. As long as this system works well, the overall health care system of a country or the world will develop and progress rapidly. Therefore, the development of artificial intelligence in China will not only promote great changes in economy, but also promote great changes in the development of the whole discipline and the development of the whole social discipline.

Attached: Academician Pan Yunhe gave the inscription "bravely entering No Man's Land" to the Hillhouse School of artificial Intelligence, Renmin University.

Hillhouse College of artificial Intelligence is a formal college under Renmin University of China, which undertakes the planning and construction of artificial intelligence disciplines, and carries out undergraduate, master's and doctoral personnel training and scientific research in this discipline and related interdisciplinary fields. The artificial Intelligence Institute carries out all kinds of work in accordance with the idea of "being able to innovate first-class systems and mechanisms, build first-class teachers, train first-class professionals, and produce first-class scientific research results". Continuous efforts will be made to build a world-class college that can influence and shape the future era of artificial intelligence. Especially for teachers and students of colleges and universities in Beijing and people from all walks of life, the college has launched a "strategically advantageous" open artificial intelligence program. The project will have two sub-columns of "Master talk" and "Youth talk". Top experts in the field of artificial intelligence at home and abroad, leading figures in the industry, outstanding young scholars and new forces in the industry will be invited to give public lectures on their focused fields. Try to gather high-end talents on the platform of Hillhouse Institute of artificial Intelligence, and issue the voice of "high-level and far-sighted" in the field of artificial intelligence.

Related articles:

Build the center of AI research in the world? Huang Tiejun: Beijing has resources, Beijing has responsibility, Beijing has hope.

Https://www.leiphone.com/news/201911/nI1uGePu9qa3xffk.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.