1. Zhang Shouyi speaks at Google: Quantum Computing, Artificial Intelligence and Blockchain;
This article was compiled by Professor Zhang Shouyi recently in Google's keynote speech.
The following is the full text of the speech:
Thank you for coming. I'm very happy to come to Google and it's equally honored to be introduced by Daya. Before that, we have been communicating with each other for a long time. Today I would like to share some personal opinions with you.
About the future information technology content, including: quantum computing, AI artificial intelligence, blockchain. Especially the connection of these three. I believe that there are many scholars in the academic community have been studying these aspects of the content, but have the opportunity to participate in I was quite excited about this academic conference because the main content of the conference research was: The internal relations among the three.
First of all, share with you an 'old' new scientific discovery.
Many scientific issues correspond to profound philosophical issues. We exist in a world of opposing unity. There are positive numbers and negative numbers in this world. There are loans and loans. There are yin and yang. There are good and bad people. The angels also have devils. There is also a corresponding reason in real life.
In 1928, one of the most famous theoretical physicists in the world, Dirac Paul, tried to apply Einstein's theory of relativity to quantum mechanics. In the process of deriving a mathematical formula, he encountered a lot of square root calculations. When you think of the square root problem in your high school period, the square root of 9 is not only equal to 3 because the square of 3 is equal to 9, and the square of -3 is also equal to 9. So when you square a positive number, you get both a positive and a negative number. .
Dirac was quite puzzled by the negative numbers drawn by the opener. This matter had a profound effect on Dirac. He predicted that all substances in the world have both positive and negative material forms. Today in Westminster In the church, you can still find a well-preserved Dirac formula manuscript.
The 'Paul Dirac Medal' obtained in 2012 was the most worthy of recognition for my research work. As I just said: When you count a number, you can get a positive and a negative two numbers. The negative number is the universe. A rule of thumb is that as long as it is a matter in the universe, there must be antiparticles. Now it seems that this is a perfectly normal perfect equation. But in 1928, in the era without antimatter, put forward such a theory. It is not easy. For example, at that time people thought that: The antiparticle of electrons is an equal mass but negatively charged particle—proton. Although the charge of protons is opposite to that of electrons, its mass is more than 2000 times that of electrons. No one believed him at all.
But what do you think he said? Lao Tzu's formula is so advanced. You guys still have to go to leveling. People really do upgrade. Dirac is very fortunate, five years later, while observing the radiation of the universe (on Earth It is difficult to observe, but it exists in the universe. Scientists have captured antiparticles and named them as positrons and electrons of the same mass but oppositely charged.
I think this is a great hypothesis in human history. The presupposition is very good, and the later results are also very forceful. Nowadays, humans have applied antiparticles to medical treatment, such as the well-known medical imaging technology--positron tomography technology. Utilizing the theory of anti-particle research, Hollywood has many similar film themes. For example, the sequel to the Da Vinci Code, which is known to all, is the climax part of angels and demons (former author Dan Brown, starring Tom Hanks). An epic battle between angels and demons to fight for anti-matter.
Antimatter can be seen everywhere in the universe. If you have antimatter or antiparticles, it's hard to overestimate your demise. But it's all based on a good corollary, but human curiosity never stops. The greatness of Dirac After the hypothesis, another great physicist stepped into our vision: Etomariano. He asked a question: Is there a substance that does not contain antiparticles, or that there is not a substance itself? His own antimatter. He made assumptions and wrote equations. But this time he was not so lucky. No one believed him. No one had ever seen the formula. He himself was also disappointed. Since then, this issue has become the basic science. One of the unsolved mysteries.
We have a list of scientific issues that need to be solved. For example, what is God particle and boson on this list? Fortunately, in the laboratory of Geneva in 2012, there was an answer to the above question. Another example is gravitational waves, its proponents. Einstein was not so fortunate as Dirac. 100 years later it was discovered that gravitational waves were a matter that was only confirmed for the first two years. This is where the magical list of antimatter particles is also found. What matter of antimatter of matter is its own problem, that is, the problem of Marianna Fermi Particles, has been at the top of this list.
Maybe looking for Marianna Fermi particles is a problem that scientists are most interested in. Like I said just now. Mariana is very disappointed because no one believes him.
Anna is Italian. But he disappeared after boarding the ferry from Napier to Palermo. This is an unsolved mystery in the scientific community. It has been 80 years since he disappeared and this year, but tells everyone that ' The good news': Not only did he disappear, but his writings were not found. This is what I focus on today. He was writing the derivation formula but he did not tell anyone where the formula was placed. This is why It took us 80 years to find the reason for it, but where did Stanford’s research team propose where and how to find these particles?
From 2010 to 2015, our team wrote three articles. The first one is to accurately predict the position of particles. What surprises us is that this particle is not in the large particle accelerator, but in our common semiconductor device. As I mentioned before, one of the substances discovered by our group ten years ago was a topological insulator, in which some magnetic dopants were placed. The materials of the topological insulator were ordinary insulators, while the magnetic The miscellaneous agent is chromium. This will form a superconductor on the surface of the two objects. We think that in this condition, we can capture Marianna Fermi particles but this is not enough. We are not only looking for The location where the particles exist needs to find the physical quantity that needs to be measured. Fortunately, common sense helps us. In general, particles have two sides, like two sides of a coin, front, back, positive, antiparticle. There are only one kind of particle in the Marianna Fermi particle: only the positive particle, there is no negative particle. He is the legendary half particle so this half particle concept will My next extremely important quantum computing speech.
Well, no matter what, the Marianna Fermi particle is a half particle of an ordinary particle. For a normal particle, there will be a certain conductivity or conductivity. This conductivity can generally be polarized by a quantum energy, showing a zero-order level two. The energy levels of ordinary particles of the isoenergetic level are discrete integers. Therefore, if the Marianna Fermi particle half particle then the energy level characteristic it exhibits should be the ordinary level of the 1/2 energy level 3/2 level. In the system of half-particles, you can measure the conductivity, but you must measure at the half-level position.
In collaboration with colleagues at the University of California, Los Angeles, last year, they turned our proposed theoretical model into reality. Based on the theoretical study, we measured the energy of the Mariannan particles. Between the two integer levels There will be 1/2 of this half-level. The 1/2-level is the key evidence that Marianna Fermi particles are half-particles, the general particle integer energy level. Marianna Fermi particle half-step energy level last year, Science magazine will After this publication was published, we celebrated well.
In that exciting moment, I remembered the movies I once watched - angels and demons. I think we found angels in heaven and did not find the devil, so we called the particles angel particles. So that's so much What's the use of traditional computers? Traditional computers are already quite powerful, but what they are good at is not comprehensive. I give two extra-large numbers to add computers, and he comes out in seconds. For example, Google Cloud, he is super fast. But if you give the computer a number, let her produce two related numbers. For example, 15 equals 5×3, but 11 cannot. You might say 11 equals 1×11, but it doesn't make sense. If you give a super super big Numbers, which are broken down into two related numbers, like decomposed 15, or like 11 cannot be decomposed, the traditional computer is afraid to be encircled. The only way is to conduct a massive search, always searching from 2 to forever .
So what is the biggest problem with computer operations? Computing like Google Cloud is essentially a process of finding the optimal solution. When the machine looks for an optimal solution, it searches for all possibilities, including finding the shortest path. It takes a lot of time to do an optimal search, which can consume a lot of time. This is why the computer needs to constantly update the iterations?
Let's take a look at the Quantum World. What is the secret of the Quantum World? Suppose there is a pair of double seams on the screen. I use a normal pistol to shoot double stitched bullets. Either through the left side or through the right side of the gap. Behind the screen You will see two bullets, one on the left and the other. But if you try to shoot two elementary particles through a double seam, the situation is not the same. It is not a situation where the left and right bullets are seen, but a complex interference pattern. When two particles pass through the double slit at the same time, the interference pattern is generated. At the same time, it passes through the two left and right slits. If not simultaneously, the interference cannot be produced. Therefore, the quantum world is a parallel world: two particles are worn at the same time. After two gaps.
Many people are thinking about the problem just now. Traditional computers have the ability to solve complex problems. Perhaps they can use a quantum computer and search for all possible answers at the same time. That is, in theory, computers can search for all possible answers at the same time. Then immediately calculate the answer for you. This can quickly increase the computer's computing power. Think of it as exciting!
I will share with you some artificial intelligence. Artificial intelligence is also a basic concept. It was proposed in the 1960s. The reason why today's artificial intelligence can make rapid progress is mainly the summation of three big trends. Iterations according to Moore's law Every 18 months can be doubled. If you use quantum computing, not only is the Law of Seoul's Law doubling in 18 months, but it is completely changed from quantitative to qualitative. Our computing power is constantly increasing, and it is almost the same as in the past 40 years. In addition, the Internet and the Internet of Things have created a large amount of big data. Big data is the best algorithm for the machine to learn, and even if the powerful computer has no data, it cannot achieve the best artificial intelligence. There is also the discovery of intelligent algorithms. , And there are rapid changes.
But the whole artificial intelligence, although everyone saw it leaps and bounds are changing, but I think it is still very early, and its future prospects are still very broad. Why do you say so? Make a simple analogy. For example, we once saw bird flying. People also want to fly very much, but early learning flights are just simple bionics. We have wings on our own arms. But this is a simple bionic, but the real realization of flying is because we understand that the first principle of flight is Aerodynamics, with mathematics and mathematical equations, can be used to design the best flight. Now the plane is flying high, fast, and good, but it's not like a bird. It's a very central point. May be artificial now. Intelligence is simply imitating human neurons, but we should think more about it. There is an opportunity for major breakthroughs in basic science. It is that we really understand the basic principles of wisdom and intelligence, basic mathematical principles, and that Can make artificial intelligence change by leaps and bounds.
In addition, what criteria do you often ask? Can you really measure artificial intelligence to meet people's standards? You may have heard of the Turing test. The Turing test is talking about people talking to the machine, but we don’t know if the back is People or machines. During the entire conversation, if you don't feel it for a day, it means that the robot seems to have reached the level of people. I don't agree with this. Although Turing is a great computer scientist, I do not approve of the verdict. People are also evolutionary processes. Many of human emotions are not rational emotions. It may not be easy for a rational robot to learn the brains of an irrational person. For example, you may intentionally infuriate a robot. Maybe it will not manage you.
So I want to put forward a new criterion. How does the machine say that one day is really beyond human intelligence? The greatest thing about people is that we can do scientific discoveries. The best criterion is that the robot can really do scientific discovery on that day. People better know scientific discoveries, and machines outnumber people on that day.
I recently wrote an article in AI that will be published in the American Academy of Sciences journal titled 'Atom2Vec', one of humanity's greatest scientific discoveries, relativity, quantum mechanics, the greatest discovery in chemistry It is the discovery of the periodic table. Today's machine assumes that we don't even know about the periodic table. Today's machine can't automatically discover the periodic table without any guidance. What we import is all that exists. The name of the scientific element, the names of these compounds are entered into this algorithm, and as a result the machine naturally discovers the periodic table of elements, which can make the greatest scientific discovery that humans think. Then our program can help us find new drugs. Use machine learning to discover new materials.
Next I will share with you my last topic blockchain. Artificial intelligence is making rapid progress. However, the most important thing for artificial intelligence is data. Today, the data is in a state of complete central monopoly, and it cannot help the machine to learn rationally. Everyone heard that a lot of Facebook personal data was stolen a week ago, at least without being allowed. In today's world, individuals generate a lot of data, personal genetic data, medical data, education data, and behavior. Data, etc. But many of these data are held in the central organization and are not truly decentralized. But the creation of a blockchain is capable of producing a data market. So in my ideal world, everyone in the future will have all their own data. This is completely decentralized storage, so that hackers are not likely to black everyone's data. Then with some encryption algorithms can really achieve in the blockchain to protect personal privacy, but also to make a good calculation.
So I describe the whole concept of the blockchain in one sentence in the future, called 'In Math we trust'. Our belief is built on mathematics. Everyone should remember this table, I see a lot of people taking pictures. One day it It will surely be the most basic of the whole blockchain and the entire IT field. It is the most basic mathematics, it can also lead to the protection of personal privacy in the data market, and it can also make reasonable statistical calculations. For example, there is a very magical The calculation method is called zero-knowledge proof. It can prove to you that my data is very valuable, but it doesn't tell you where the real privacy data is.
The topic I reported today is mainly a core concept. It is to make IT really developable. It requires both physics and mathematics. Shenzhen has done a good job in applications, but because the university is not the world leader University, but I suggest that the most core investment, this type of mathematics and physics, is really closely linked with the IT field.
With the blockchain, the emergence of this data market, we can really make society more fair. The biggest injustice in our society now is that we easily discriminate against a few minorities. But what is needed most in the process of machine learning is Data owned by those minorities. If the accuracy of today's machine learning reaches 90%, I want to achieve 90% of the 90%. It needs not the data that has been learned, but the most different data from the past. It is the minority-owned data that are most valuable for machine learning. Once we build on the blockchain and add these wonderful mathematical algorithms, we can truly reach the data market in this data market. Inside, the data owned by these minorities is the most valuable. In this way, we can really turn an ugly duckling into a beautiful swan. Because ugly ducklings are not ugly, they are just different from others. In this world, we really reach a blockchain. The world concept of coexistence with artificial intelligence, they will be the most valuable.
Throughout the blockchain, everyone’s understanding of it is not yet the most fundamental first-principle understanding. Using the most basic principles of physics, to reach consensus is that everyone agrees on the same book, which is equivalent to physics. For example, magnets are inherently disorganized, but in the ferromagnetic state, they all point in the same direction. Therefore, reaching a consensus exists in the natural world, and in today’s human world. However, this phenomenon is a phenomenon called entropy reduction. When we reach a consensus that everyone is heading in one direction, the entropy of this state is much smaller than that of chaotic entropy. It is very difficult to reach this consensus, because the entropy is always increasing, and it is very difficult to reduce it today. The thing that can achieve a consensus system in the blockchain is to use an algorithm that consumes energy on top of it. You may not understand why this thing sounds unreasonable at first and why some accounts consume energy. From the second theorem of physics, this is a very reasonable thing, because reaching a consensus is itself an entropy reduction, but the entropy of the entire world must be increased. Consensus at the same time must take some additional entropy excluded. This is no centralized mechanism with the natural world inside the magnet from the chaotic state to achieve an ordered ferromagnetic state is very similar, which is the inevitable price to pay trend.
I am here to share with you that apart from being a professor at Stanford University, I am also the founder of Danhua Capital. Our main core philosophy is to connect today's most cutting-edge technologies and investments, and to use first principles. Way of thinking to understand the world today.
The other thing I want to say is that I come from the academic world. We need to make two big bridges in the whole artificial intelligence field. One is to have a close connection between the academic world and the industry, the best physics in the academic world, the best mathematics. And the discovery and invention of the algorithm. On January 8th this year, I was very honored to receive the People’s Republic of China International Science and Technology Cooperation Award granted to me by the General Secretary Xi’s. Our entire world of science is the most endless, with no national boundaries. Yes, science can truly bring human beings beyond national borders. The artificial intelligence that we need to solve today, quantum computing is the problem of the entire human race. So we must not put our eyes on our own parts, but we must look at the whole world. And the whole world. In this process, China is also a very big opportunity. Everyone wants to answer the question. In addition to applying well-developed technologies to China, can we have genuine original technology in China. I’ll introduce you today. These are the basic physics and the most basic mathematical principles. We can do this well, and these principles sound better than Abstract, such as the principle of entropy increase, electron-positron, but at the most basic level, this is our wonderful world of today, it can really provide development prospects of science and technology essential to the entire IT industry. Thank you.
Need I answer some questions?
'So you mentioned consensus, and the work proved the system reached consensus through dispersion and by increasing the unity.'
'Yes. '
'How did it get in the space unified system?'
'Okay, actually, I think there will always be some trade-offs in the end. I think the future of blockchains and those password currencies will be similar to our current world. The world today will have M0, M1, M2, different levels, So I think the most basic layer. The global currency should be completely based on proof of work. Because the unity after which you dump is completely transparent. Not only because it must be there, but it must also be completely Transparent. I think the most basic and fundamental level, the status proves not to be useful, because there is too much possibility of collusion. You will lose some of the linked things, but also get some unlinked things, you will get bribery. Behavior, so I think the most exciting thing in the blockchain world.
It is on the most fundamental level that there is something completely objective, and that it is only connected with the real world, that is, energy. It is not connected with the state proof, otherwise it will be filled with human irrationality. But I can imagine that in the They will be useful at higher levels, but at the most basic level. M1 or M0 needs to be completely stable.
So I still think that work proves, or there is another method called space time proof, space proof. Based on capacity, I think it is also a kind of quantifiable physical resource. I think that is the least human being that should be included. Something, but maybe it is the most included.
I usually think that quantum computing is a useful search algorithm for artificial intelligence, so one of the most interesting methods of artificial intelligence. So I'm not saying that these three trends must work together forever, in fact they can progress through competition. .
So on the one hand, quantum computing and blockchain are competing with each other, because the creator coding algorithm may be defeated by Compton. On the other hand, I also see that Compton can help artificial intelligence to do the most efficient search. It's also what artificial intelligence needs to do. So this relationship is very much like the symbiotic relationship in our ecological environment.
We can't assume with human thoughts that they will always do the same thing. I think that in the course of competition, they will all become stronger.
'You mentioned the global currency or M0 or 1, I'm curious, I know you're a theoretical physicist, but in the implementation process, or take the iPhone for example: My iPhone7, I met iPhone6, I met the iPhone5, But there is a consensus of a metal layer that needs to be achieved, so I actually agree with this decentralized system.
'Correct. '
'There are currently many fragmented pools in cryptography. The evidence of liquidity has not been quoted. So how did you bridge the gap we are in?'
For example, the relationship between the Bitcoin blockchain and the Lightning Network is very suitable for the M1 M2 framework, so basically the blockchain is completely objective under proof of work, so it tries to achieve the best among the parties. World consensus. The parties do not know each other, but they still need negotiation and exchange. But when you really think about business affairs, maybe we both have been happy to cooperate as partners in the past ten years. Then why should we still be when the other party is What can we do with strangers? So what we can do is to step into each other's status channel and put our Cradow on the blockchain, but we still do very, very fast transactions, but we still pick one load per month. I think this is like the relationship between M0, M1, M2, the relationship between lightning and Bitcoin, just like the relationship between M0 and M1.
So when you go up one level, stability will decrease, efficiency will increase, but the tradeoff comes from our history, we already have a history of trust, so if you have business partners, they are familiar with each other and they don’t need to use the largest number at all. The stable layer. They can build a higher level, sacrifice some universality, but in return it is an increase in efficiency. '
'I have a question about angel particles'
'Ok. '
'Angel particles are not positive?'
'Yes, it's half the bit.'
'Sounds like an authentication element in the real world, when faced with their own problems...'
'No, in a more precise analogy it's like a bunch of complex numbers can be expressed as two real numbers, so a complex number is like a particle complex paired like a repulse particle. If you have a real number, Complex pairs are just like themselves, so angel particles are more like a real number.
'I understand. We think this is the contrast between yin and yang.'
'Yin and Yang, right.'
'What would be the middle element, what would be an angel...?'
'I think the analogy is just saying that there is an upcoming qubit here, but in actual calculations, before you do actual calculations, you are splitting them, but through splitting, they are already non-local, they are tangled together, But classic interferers are not confused, so they cannot be removed with classical interferers. This is why the topology quantum computer is so stable.
'Yes. '
'Okay, so combine several topics of your speech: If you can manage the energy of quantum computers, then if we can securely manage our data, we can share data through privacy encryption.'
'correct. '
'I want to know what you think about Google's future? Because it's like a real threat. If anyone can get into a quantum computer doing efficient search operations.'
'Yes. '
'Then they can control anyone's data.'
'I think the only way is not to resist change, but to embrace change.'
'correct. '
'So how do you see that Google is mastering this technology in the future world?'
Actually, I have an answer: So in fact, we can carry out the following ideas. For example, my personal data, I want to store it in a safe way, but I can still do some calculations. We know that Google Cloud Disk is Compete with Amazon Cloud Disk, so what we can do is to store completely random numbers in Amazon Cloud Disk, but in Google Cloud Disk I store my personal information, plus the random information I stored in Amazon Cloud Disk, so if I It is impossible to predict that these two highly competitive entities will not colluded. They cannot exchange information. Then you can use the system of multi-party security calculations to do calculations. Get a calculation result without leaking any details. So in this world, centralized entities are still useful. But to make it work, you must assume that the two sides are competing, not colluding.
'Hello, I think the use of terminology uniformity is interesting because it seems to be a mysterious thing, but it is also very valuable. In thermodynamics you can have logarithms in classical thermodynamics and then you have a unity The theory of information, and then you make an analogy with energy, which reminds me of some free energy.
Yes, that's true. I think the blockchain world does extract some free energy from it, so you basically got something, but no matter what amount of energy you got, the useful amount, the energy you use minus Go to the unity you have to waste. So today you can still see a lot of white people. They claim to be miracles. Such white people reminds me of the pursuit of perpetual motion in the eighteenth century.
'I want to know you can further speculate analogy, you need a temperature to make it work.'
Right vs. True. In fact, the temperature appears very natural. When you have a saving amount, such as energy conservation, the temperature concept naturally appears. Because whenever you have a random but economical system, this is the most common , is called the Boltzmann distribution. So the introduction of temperature is natural, but why I am so excited about this is because this is the first time I have seen the confluence of social sciences and natural sciences, it provides the world of social sciences Rely on, so I think M0, M1, M2, as the basic anchor.
Now that we have entered the natural sciences, we can see exactly the unity, it is wasted. So we can know why it reached a consensus, and then you build on it more human things. But at the most basic level, It is now a consensus between society and natural sciences that fundamentally reduces energy, unity and information.
'Thank you very much for spending so much time. I think in your speech, you are saying: You will see the first layer of a blockchain, and then build more layers on this basis, so you feel a wide variety of items And companies are trying to build their own blockchain, what does this have to do with your speech?'
'I think that you provide something unique. So bitcoin, blockchain and theorems vary a lot, so because you're the foundation of trust, you don't want a universal Turing machine. Because maybe it will be black , But you have to do more processing on it, and then if the theorem looks more natural, the evolution of the blockchain world will be very similar to the evolution of biological species. You see branches, different species, if they Forking long enough, perhaps they become different species, but always have the most basic things. All living things are composed of cells, so this basic contract will not change. But for the organization, different Organisms are composed of different cells and they may change.
'Yes. Thanks for your explanation.'
'My question is: When do you think quantum computing can actually be applied? After your discovery and research comes out, and then when it is actually applied, do you think it will only be grasped by a particular large company?'
“Good and good. I think quantum computing research should ideally be open to use. I think, let me declare: I know that many companies are trying, but the basic principle of corporatization is that they want to protect the interests of shareholders. They want to protect their secrets, but for this energy-heavy thing that doesn’t know anything about human influence, I think it’s better to do it in open university studies, and that’s what I’m doing. .
So I'm dealing with quantum computer methods. There are very, very many companies trying to join my quantum computer research, but I'm against this.
'What are your predictions for the use of quantum computers?'
'Does it contain or contain my invention? I think if you try it this way, it will take a lot of time, can you imagine? You need seventy bits to serve a useful bit, I don't think it's worth it. But with This method is worth it.'
'Okay, I think we're almost over, I asked the last question: About your angel Fermi, did it change any other requirements for quantum computing, such as absolute zero?'
'No matter no, it's still running, it's running at low temperatures most of the time, but unfortunately our method can be run at room temperature, if we find a superconductor that can run at room temperature. If there are real improvements, we can It drops to low temperatures.'
'Thank you very much, Professor Zhang.' Phoenix Network Technology
2. Ministry of Education: 71 colleges and universities have set up artificial intelligence related disciplines;
Xinhua News Agency, Hangzhou, June 8 (Reporter Yu Jingjing) The Ministry of Education recently issued the “Artificial Intelligence Innovation Action Plan for Colleges and Universities”, which proposes that Chinese universities should be divided into 'three steps' and become the core force of building the world's major artificial intelligence innovation centers in 2030. And Talent Heights.
On the 8th, the Ministry of Education held a press conference at Zhejiang University in Hangzhou to interpret the Action Plan for Artificial Intelligence Innovation in Colleges and Universities (hereinafter referred to as the “Action Plan”). The director of the Department of Science and Technology of the Ministry of Education, Lei Chaozi, said: Focus on and strengthen the basic research and key technologies of the new generation of artificial intelligence, focusing on big data-driven knowledge learning, cross-media collaborative processing, human-computer collaborative enhancement intelligence, group integration intelligence, autonomous intelligent systems and other directions.
Lei Chaozi said that according to the “Action Plan”, the development of artificial intelligence in universities will be divided into three steps: First, by 2020, it will basically complete the optimization of the scientific and technological innovation system and discipline system for universities in order to adapt to the development of new generation of artificial intelligence; By 2025, colleges and universities in the field of new generation artificial intelligence will significantly improve their ability in scientific and technological innovation and personnel training, and obtain a group of original achievements with internationally important influences, effectively supporting China’s industrial upgrading, economic restructuring, and smart society construction; In the past year, colleges and universities became the core strength of the construction of the world's major artificial intelligence innovation centers and the talented highland leading the development of the new generation of artificial intelligence, providing scientific and technological support and talent protection for China's advanced innovation-oriented nation.
The Action Plan also proposes the 'implementation of' artificial intelligence + 'actions' to support universities in the fields of smart education, smart manufacturing, smart medical care, smart cities, smart agriculture, smart finance, smart justice, and national defense security. In particular, we must promote the development of smart education.
'Artificial intelligence is a strategic technology that will lead the future. We must truly understand and grasp the new opportunities for the development of artificial intelligence.' Academician of the Chinese Academy of Engineering and former vice president of the Chinese Academy of Engineering, Pan Yunhe, said that the Ministry of Education and the Academy of Engineering will soon carry out consultations in the field of intelligent education. The project will provide advice on future education paradigm changes. Personalized education based on big data intelligence, cross-media learning, and life-long learning will promote changes in educational goals and concepts, and even reform and innovation of the entire education system.
It is understood that the Ministry of Education has now agreed to support the establishment of an artificial intelligence collaborative innovation center at Zhejiang University to accelerate the establishment of a highland for scientific and technological innovation and personnel training in the field of artificial intelligence in China. Zhejiang University President Wu Zhaohui said that Zhejiang University will focus on machine learning algorithms and Big Data Intelligence. , Cross-media awareness computing, hybrid enhanced intelligence, human-computer collaborative intelligence and other major scientific frontier issues, accelerate the establishment of artificial intelligence first-mover advantage, and serve national strategic needs.
According to reports, as of December 2017, a total of 71 colleges nationwide had set up 86 secondary disciplines or interdisciplines in the field of artificial intelligence. The Ministry of Education identified the first batch of 612 'New Engineering' research and practice projects in 2018, including the layout 57 personal intelligence projects.
3. Huaying Capital Ji Wei: 2018 is a period of no investment;
On June 9th, in 2018, Hua Ying Capital Annual General Meeting was held in Shanghai. Investors, representatives of invested companies, industry experts and founding managing partner of Hua Ying Capital Ji Wei, partner Liu Xianming and other senior executives attended the forum together. The nine-year investment history.
As China's earliest cultural industry fund, Huaying earlier saw the integrated business model of culture + consumption, technology + industry, etc., and fully opened the 'content +' layout. At present, it is the 'head investment brand' in the cultural field.
The theme of the conference was 'Ideal Sansuan'. The investment companies in various industries such as consumption, entertainment, content, etc., shared their respective fields. Liu Xianming, partner of Huaying Capital, said that when the era of the mobile dividend is almost over, the data will become the first 1. Production data, technological innovation is the engine that drives the next round of commercial transformation.
Through the “China Investment Capital Annual Investment Report”, the founding managing partner of Huaying Capital, Ji Wei, answered the question of how to look for convincing value and seize the high-quality entrepreneurial project with both ability and ambition.
Ji Wei said that in the past year, frequent enthusiasm, over-inflicted hot spots have become difficult, 2018 is the year of financial supervision, and the introduction of the "New Assets Management Regulations" will have many restrictions and restrictions on funds. However, from the data point of view, the investment in Q1 in 2018 was a new high. This was due to the fact that the capital resources were closer to the head of the project, and 1% of the investment cases accounted for 30% of the investment amount.
Culture and entertainment investment in the consumer sector needs long-term brand management. It cannot be accomplished overnight. It is not a tuyoke-type investment. Gradually fierce artificial intelligence, blockchains have been gradually used in security, finance, and retail industries after a long period of development. And the project landed, but these hot spots also need long-term training, long-term market progress can promote its growth.
Based on this, Ji Wei said that Citi Capital believes that it is now an investment period with no outlet, and it is also an era where capital can favor long-term and value return opportunities.
Ji Wei predicts that in the next decade, China will usher in the golden decade of brand rise. From the perspective of consumers, the maturity of the entrepreneurial team, or from the infrastructure, upstream production, supply chain to downstream payment, and logistics foundation, have been very Complete, the brand market has reached a stage of rapid outbreak.
She emphasized that the two investment drivers that Huaying insists are the changes in user needs and technological innovations. Professional competence, market ability and operational capability are the three core competencies developed by Huaying.
In addition, at the conference site, Ji Wei shared her belief in the field of venture capital: 'If the venture is a magnificent journey of 60 months, then it is likely to be the 'waves' of the first 59 months and the 'magnificence' of 1 month. It is only possible to usher in the light of the last business.
Since its establishment in 2008, Huaying Capital has invested in nearly 150 companies and has successfully listed or cancelled more than 20 projects. Among the invested projects, there have been more than 100 cases of subsequent rounds of financing, of which 26 were valued at 500 million yuan. Nearly two-thirds of the invested companies confirmed the completion of the subsequent round of financing, with a total financing amount of over 6 billion yuan.
4. Prediction of artificial intelligence in Russia: Germany won the 2018 World Cup and is difficult to qualify in Russia;
China News Service reported on June 9 that according to the report of Russian Satellite Network, Perm National Research University, Russia, the artificial intelligence (AI) produced by the students predicted that the German team will win the World Cup in 2018.
5.AI medical usher in the Chinese era
AI Medical Arrives in China
Source: Economic Observer
Ni Hao
The future of artificial intelligence is medical care, and China has world-class propositions. Propositions determine the height of scientific work, and strong demand can boost the development of technology. The origin of medicine is transformed into computer again. Ni Hao, who has a cross-disciplinary background, knows well that the medical industry Requires artificial intelligence, and the Chinese scene has great potential.
In January-November 2017, the number of medical emergencies in medical institutions nationwide exceeded 7.3 billion, with an average of over 21 million visits per day, and it is still in continuous growth. Coupled with the uneven distribution of quality medical resources, the development of regional medical care varies greatly, only AI empowers medical institutions to further improve the operational efficiency of top medical institutions, amplify the quality medical resource service capabilities, improve the level of medical treatment at the primary level, and promote the homogenization of disease diagnosis and treatment, so as to realize the development goals of 'common prosperity and national health'.
China's AI Medical World Proposition
At present, medical artificial intelligence related companies can be roughly divided into three categories. The first category is such as Tencent, Ali's Internet company; the second category is a multinational company such as Philips, Siemens; the third category is the start of the information system manufacturers. The medical care you want to do is different from these three types of companies.
If the enterprise is doing medical treatment based on platform thinking, for example, collecting data from hospitals and uploading videos to the cloud to get results, there is not much combination of this practice and clinical practice. However, the user experience emphasized by the plans is how to make users second Get the system's response results, obviously platform approach can not achieve this goal.
Multinational corporations involve differences in the environment. As the global model market, most of the top R&D personnel of medical international giants are in the United States. Nowadays, the advantage of Chinese companies in the development of medical artificial intelligence lies in the fact that the differences in the medical industries in China and the United States themselves determine China. The demand for medical artificial intelligence is much stronger than in the United States.
First, the problem of insufficient medical imbalance in the United States is relatively weak. The differences among doctors are very small. Basically, they all have to receive more than 15 years of education, and the general practitioner system is well developed. The differences in medical standards between urban and rural areas are not. The gap between doctors in China's county-level hospitals and doctors in Shanghai's top-three hospitals is so great. Moreover, American doctors respect clinical medical guidelines, the standardization of diagnosis and treatment is relatively high, and homogenization of medical services is better. There is a gap in this aspect in China.
Second, the demand for medical artificial intelligence in the United States is not as urgent as in China. According to World Bank data, China's total health expenditure accounted for 6% of GDP in 2015, while the United States maintained over 17% year-round. The United States' GDP in 2015 was 17.4 trillion US dollars. Converted into RMB, more than 300 million people spent RMB 20 trillion in medical expenses. In that year, China’s 1.4 billion people spent 4 trillion RMB in medical treatment. It is undeniable that the Chinese government’s investment in healthcare continues to grow. However, the gap is considerable. The overall investment in medical resources in the United States is already sufficient. The need to use artificial intelligence to solve problems seems less urgent.
Third, the human cost of developing medical artificial intelligence in the United States is relatively high. Medical artificial intelligence must work collaboratively with experts in the medical field. American doctors are the tip of the social income pyramid, with an average annual salary of more than $250,000. In contrast, there are only 8 computer practitioners. Million dollars, technology companies can hardly afford a large number of high-level doctors to develop products together.
Therefore, China's medical industry has world-class propositions. Regardless of the depth and breadth of technology or clinical application, China's medical artificial intelligence has taken the lead in the world. Its greatest advantage comes from the country’s booming growth rate, coupled with artificial Intelligence has been promoted to the national strategic level, not just as an applied innovation. From the perspective of practitioners, the Chinese era of medical artificial intelligence will certainly emerge.
AI medical tentacles
The problem of inadequate social development in China at this stage is also reflected in the medical field.
In the area of oncology, for example, the country is strongly encouraging high-risk cancer patients to undergo early cancer screening. However, hospital capacity that cannot be rapidly increased cannot meet the large demand caused by policy promotion. Therefore, there is a shortage of doctor resources. In addition, the top three hospitals. There is a wide gap between the overall strength of doctors and basic level hospitals, and the level of equipment, etc. There are often cases of false positives. If a patient checks at a primary hospital, the conclusion of cancer is reached, but after checking at the top three hospitals The conclusions may be completely different. Imbalances and inadequate problems also occur in the pediatric field. The lack of pediatrician resources has led to many hospitals cancelling paediatrics, or temporarily deploying other department doctors. When they studied in the medical school, almost nobody wanted to choose pediatrics. .
In accordance with the plan is one of the earliest enterprises to carry out medical artificial intelligence layout, there are two priorities in the selection direction, first, to see how high the demand in the medical field, and second, to see how much clinical value. According to the plan to select the tumor, heart Cerebrovascular disease and paediatrics, which are currently the three areas with the greatest demand for artificial intelligence.
Since the introduction of medical care in 2016, there has been a shift in the mindset of the plan. The initial choice was for lung cancer screening. After the screening process was done well, it began to consider whether it could be replicated in other areas. For example, blood tests or After filming, the doctor needs to give a specific conclusion. In this link, the difference between doctors' conclusions is more obvious. Not only that, the clinical demand for artificial intelligence is equally huge. Therefore, the idea of following the plan serves the purpose of serving the radiology department. The Department of Ultrasound and Laboratory Branches are transformed into clinical departments for layout services, and provide artificial intelligence products for doctors in the Department of Respiratory Medicine, Hepatology, and Oncology. In addition, according to the plan, “urban radiation in rural areas” will be used before landing in top hospitals. This year, the system will be pushed to the grassroots.
AI Medical Big Data
It is based on a deep understanding of the industry, according to the map often mentioned 'talk about the algorithm is an outsider's performance, the experts talk about standards and data'. And the understanding and processing of the data is based on the figure can put the artificial intelligence landing clinical core Competitiveness.
To understand medical big data is the most basic and important. It is not to say that centralizing medical data in various places is called medical big data. This can only become 'large medical data'. The essence of big data is to support data from it. Excavations and insights.
For complex health data such as clinical data, population health data, and public health data, artificial intelligence can exert its powerful ability to analyze complex text data, reduce the dimensionality and structure of medical image data, and purify the quality analysis and control. , Diversification of data standardization capabilities.
Pick out the sand from the rice, wash the dust on the leaves, and use appropriate temperatures and superb cooking techniques to bring together the unorganized raw materials to order an orderly and pleasing medical big data feast.
When medical big data is understood and available, the next step will be to establish a rapid replication capability. One class of disease development products can be migrated to another type of disease to speed up development.
In accordance with the map, it has been hoped that the country can establish a health big data alliance to open up the medical big data. Although China has developed an electronic medical record after 2010, each patient can find his own case in the hospital. But the biggest problem is that there is no Get through, see the disease in this hospital can be found, not a change of one.
Opening data requires centralized storage of data. Now, there is a data exchange in Guizhou. If the Exchange can open up the data of desensitization of the Health and Planning Commission, such as the number of lung cancer patients in Guizhou, regional distribution, disease progression, etc., this will help Benefit medical big data research.
After completing the analysis and processing of the data, artificial intelligence will also exert powerful scene-based application construction capabilities, such as the preparation of clinical research data in all dimensions, the improvement of the standardization of diagnosis and treatment, and the expansion of the supply of basic medical services in China, and will continue to improve the medical institutions. The clinical services and scientific research capabilities were eventually built into smart hospitals.
(The author is the president of Yitu Medical, the newspaper reporter Chen Yifan finishing)
link
As a unicorn enterprise, founded in
Int'l Technology in 2012, its people
The technology of industrial intelligence has been used in many fields such as medical and health, urban public safety, and finance. In 2016, Intu Technologies entered the medical field and became the only one to deploy medical care.