Hottest AI engineering doctor Zhou Jin AI and bloc

AI engineering doctor Zhou Jin: AI and blockchain technology

Golden Finance - blockchain news on May 21: Yang Binyu, on May 21, 2018 Fanzhi blockchain summit · Tokyo was grandly held at the Grand Prince Hotel in Tokyo, Japan

Fanzhi summit is one of the most influential blockchain summits in the world, which has been successfully held in Shanghai and Macao. This Tokyo summit is the third stop of Fanzhi summit in the world. It brings together top projects in the blockchain industry, excellent enterprises, government officials, investment institutions, scholars and various media, aiming to strengthen industry exchanges and promote the rapid development of global financial technology

in addition to sharing wonderful blockchain projects, the event will also discuss topics such as the Japanese government's development support plan for the blockchain industry, exchange compliance, and the regulatory direction of cryptocurrency

Zhou Jin, Ph.D. in artificial intelligence and head of Huajin International Ai division, delivered a speech on the theme of "artificial intelligence and blockchain technology"

the following is the actual record of the speech

Zhou Jin, Ph.D. of artificial intelligence engineering and head of Huajin International Ai business department

good afternoon, everyone. I think the guest's speech this morning should be fruitful. I don't know how you feel? At least I still feel that there are quite a lot of undigested and cutting-edge technologies. Today, I am very happy to have the opportunity to share with you the cutting-edge technologies, artificial intelligence and blockchain, because I want to share the topic with you from the perspective of a technology researcher. Let's look at it in a very easy to understand way. I am a professional in artificial intelligence. First of all, let's make a brief self introduction. My name is Zhou Jin, who graduated from Waseda University and majored in artificial intelligence, At that time, my research topic was a branch of artificial intelligence, that is, evolutionary algorithm in elevator group control management and application. At that time, the subject cooperated with elevator and escalator manufacturers

I want to share a topic with you about artificial intelligence. As for the past and present life of artificial intelligence written above, you may ask, why is it the past and present life? As we all know, artificial intelligence has been out for many years, but it has been very hot recently. Everyone says that 2017 is the time of artificial intelligence, and 2018 is the time of blockchain, so I'm here to have a look at the past and present lives of artificial intelligence

the background shows the development history of artificial intelligence, so I'll simply go over the situation of artificial intelligence with you here. From the background curve, we can see that artificial intelligence has actually experienced several stages of development. First of all, it was put forward at Dartmouth society around 1956, and then it went through several stages of development. I think you can have some impressions. In the early stage, it was from analog control, and I think you may still have some impressions. Then after that, some evolution has been developed, and this kind of model has been invented by combining the advantages of xinyilian platform of China 5 mining and Alibaba's 1688 channel. With years of technology accumulation, gap and so on, so our research topic was also called this, called the application of evolutionary algorithms. In fact, this algorithm refers to the theory of evolution. In fact, it has crossover. In fact, it is also a process of simulating the theory of evolution. So it realizes crossover application in the model as its core algorithm

in addition, he is actually about neural collaterals, which are also called artificial neural collaterals. To do research from another perspective, it refers to the neural collaterals of the human brain, to follow the core parts outside him, through modeling, and through adjusting the structure and parameters of the collaterals to integrate the input and output. In fact, there are two main problems in the neural network. One is that it only has output, but it can't see the process, because the whole design idea is to do the sum of input and output. The other is over learning. What does it mean to be over learning? You say that your training effect is good, but when you apply it, you will find that the actual effect is not very good. Now these problems have been solved to a certain extent, because there is a large amount of data in the area, as well as the structure of the whole network

then, in the history of development, I want to talk about that in 1997, IBM developed the deep blue II supercomputer, but what caused a sensation was the victory over the chess master. We can look back on the career of computer. Then I think everyone should be familiar with it. In 2016, ahphago, which has been considered to be difficult to overcome in this field, that is to say, after ahphago was developed, it defeated the world champion of go, This time, I think it should be known as artificial intelligence, deep learning, etc. after that, it actually goes further. So far, all computers are developed based on models. In 2016, the development of the so-called new concept of neurons, how is it IBM? In fact, IBM everyone knows that deep blue, deep blue II, it is still in-depth research in this area

it is said that there are a series of neurons. This is an impression map, that is, this 64 million neurons, 161 nerve synapses, which actually simulate the structure of human beings. It should be said that it is a very revolutionary stage in the development history of artificial intelligence. On this basis, the popularity of artificial intelligence is getting higher and higher. In fact, I think you are familiar with Hawking's most famous prediction, that is, in all fields of artificial intelligence at present, its application effect and actual performance, those just mentioned, and even the development of neurons, In fact, of course, Mr. Hawking has made some definitions of artificial intelligence, which he studied in 2017. It may be seen from the current development that AI has a great impact on all aspects of human life, work and life, including the real work and business itself. After all, what will happen in the future? In fact, the direction of technological development is also a matter of great concern to everyone

then let's make a reply to artificial intelligence. I have done artificial intelligence for so many years. I have a team and Research Institute to share with you how to objectively look at artificial intelligence and its future trend

first of all, I think there are often several problems between artificial intelligence and human beings. What are they? For example, I think artificial intelligence is actually evaluated by people as IP, but in fact, as human beings, in addition to IP, there is the so-called EQ, and we have not even heard of it, which shows that when we compare artificial intelligence with human beings, we actually ignore or do not realize that there is a difference in the comparability of comparison. In fact, in these fields, compared with human beings, it actually shows the very advantage of artificial intelligence. In fact, looking back, they are not comparable. In fact, in this morning's speech, the gentleman in Singapore mentioned penetration, But in fact, the field of artificial intelligence is also trying to develop creative systems and robots, but in essence, there are some big differences with human creativity. In the abstract, I want to say that the sixth sense of human beings is in the process of human thinking, and artificial intelligence is in the so-called operation process. In fact, its method is different from human thinking, which is abstract. As a human thinking can be confused for a while, and human beings have sudden inspiration. What does this mean? It's actually creativity

next, I want to tell you about these successful cases, such as ahphago. About ahphago, let me share with you. In fact, ahphag is small. O itself has undergone an evolutionary development process, which is quite fast. After coming out of the first generation of ahphago, it has caused a sensation in the world, the ahphago generation. Then it developed a new generation of ahphago. In fact, we are looking at it specifically from researchers? In this AI product, we will find that in fact, there is a core algorithm here, which is the so-called search algorithm. In fact, it is not a simulated human algorithm. Its algorithm is more suitable for AI. In the process of fighting with people, we will think it is a comparative logic, but in fact, after our analysis and in-depth research, we find that it is actually different from people, That is, board games may be effective. In fact, it may be used in other, maybe other parts of the game. For example, you can apply it in the financial field, but it may not work, that is to say, its homology is not good. This mentioned about ahphago

there is no doubt that when I was a doctor, if there was a need for some breakthrough research in the field of neural networks, there were actually limitations in the existing architecture. At that time, I thought that unless neuron deception was mentioned, it was actually developed two years ago after seven or eight years. This is still very necessary. The development of this technology is really unpredictable, Let's take a specific look at the developed ones. One is to simulate the human brain. In the process of simulating the human brain, compared with the physical structure of the human brain, the neuronal synapses are actually different, and the neural synapses are actually different in quantity. As far as artificial intelligence is concerned, so far I think human beings have no need to worry about the uncontrollable development of artificial intelligence. Do we want to suppress this development? In fact, from another perspective, why are humans more intelligent than AI? It is precisely because human beings have talented scientists like Hawking, who can prepare for a rainy day. I think artificial intelligence is very difficult to do

I think instead of worrying about these topics, I'd better further develop the concept of AI. AI is being widely used in some aspects and will be applied in the future. First, in the field of finance, this is one of the themes of today's conference. This is actually a product developed a few years ago. JPMorgan Chase's financial software. In fact, we can see that its work efficiency is completely different. Once again, I want to say that this can only say that we human beings develop it better. This is a means, and we can fully enjoy the dividends brought by our development of this invention

second, another financial service provider, its work efficiency is actually very amazing, that is to say, in the future, this kind of financial technology talent will actually further become the demand of this era. We are engaged in the financial technology industry and the radiated field of service. If you enjoy the dividends brought by AI here. That is to say, there are actually several developments, and I won't explain them in detail in the later stage

another area is about autonomous driving. I want to go deeper. Google and Baidu actually have a lot of industries. It shows the current achievements of major automobile manufacturers and technology companies in this field