Every day, 30 million cars are processed.

category:Internet
 Every day, 30 million cars are processed.


[NetEase intelligence April 27th] 2018GMIC the global mobile Internet Conference opened in Beijing yesterday. The theme of this conference is AI is everything. At the forum, a speech was made by the head of the American Institute of research and vice president of the security strategy, the vice president of the security strategy, on how to use artificial intelligence to improve travel experience, especially to ensure travel safety. Gong Feng Min said that one of the vision of drops is to transform traffic with big data and AI technology, improve travel, environment, ecology and life, and emphasize safety. Technological transformation of transportation is a global challenge. Security is the key to the success of this challenge. Gong Feng Min said that the following measures are used to ensure the safety of passengers and drivers. Trickle trip platform needs to complete 30 million single taxi needs every day, which generates a lot of travel data. Feng Feng min believes that first, we must ensure data security, pay attention to user privacy protection and policy compliance, and focus on developing data leakage awareness. Second is driving safety, isomorphism to track the track and analysis of users ecological behavior, so as to ensure the safety of passengers and drivers. The third is the safety of the Internet of things. As a travel tool, the car develops from the traditional car to the auto driving car. Besides the physical safety, it should also do the vehicle communication, the human vehicle communication, even the whole intelligent transportation system, and consider the whole environment and the ecological problem, and solve the security problem well. Facing the problem of man and car, Gong Feng min revealed that the current drops are taking several means to combat the inconsistency of the human vehicle, including intelligent interactive blocking, multi behavior face recognition, one key report complaint channel, human vehicle photo and manual examination and so on. For the relationship between artificial intelligence and human beings, Gong Feng Min said: what people need is not to defeat their opponents, but to be an assistant. (Yi) The following is the bow peak speech: Gong Feng Min: Good afternoon, everybody. Its a great honor to come to the GMIC meeting. I used to be a creative person for decades, and then I went to the Internet field to have a greater impact and to touch everyones life. So today we share with you more about AIs sharing in the travel area. The main introduction of the drop mission, the most important is to solve from AI, to help us make a better trip a few dimensions, from the security experience to the efficiency and efficiency to share with you. Obviously from the point of view, many of us know more about it. In fact, a few very important missions, the first one, we think that with big data and AI, we can change traffic, making it a traffic is a global thing. Second, we start with improving our travel, and then we can change our environment, change our ecology, and make our lives better. This is also our hope that it can affect everyone. Third, we realize that it is a matter of technological change reform. It is important to be safe or first. Besides security, it affects our lives, and the new things are often accepted by everyone. There will be a lot of noise when we do it. At this time, it is possible to say that security is also possible. If we fail to do well, it may affect one of the factors that we stick to our right path. So safety first, experience second, then efficiency. This is a basic idea for us to do things. Last but not least, we realize that this is a great challenge. We must do well in terms of cooperation between ecology and policy. Today, the theme is related to AI. We all know that AI is the AI of Turing thinking. At that time, it was said that machines could not do human thinking, but the starting point was very low. Later, from Alfa dog many successful cases, in fact, we are thinking about the concept of machine intelligence. At this time, we say machine intelligence is that it can do something, we realize that people can not actually do it. I will not stay on this point, so the next important turning point is from machine intelligence to intelligent machine. Although it seems to be a single word, but in fact we not only realize its ability at this time, it does a complex job, but also because the machine itself can be superhuman in many aspects of the physics, so at this time we regard it as the most effective tool, which is also a core concept. This is not much to say. The important conditions behind these two are machine learning algorithms, and I am not an expert. The other one, big data. So the combination of these two and powerful computing power will enable us to do a lot of things that AI cant imagine before. From the point of view, we also know that because we want to solve the problem of peoples travel, we have different needs at the beginning, and we may have a series of different services, which may have different needs and different scenarios with different people. On this basis, we need to solve such a service. If we want to get thirty million of our services every day, in fact, it involves a lot of data calculation. I dont talk about these simple numbers. Just take a look at it, such as completing the thirty million single services per day. In fact, the calculation of the various path requests we need to do is 40 billion. Then we went to see the amount of data processed was very, very huge. On this basis, this is the opportunity that we use AI technology to do better, do better in efficiency, and do better in experience. I look at it from three aspects: first, from the perspective of security, to see what kind of application scenario it is. In fact, there is a very interesting thing to do here, because in the early days when you were always at the time of being safe, what was his means, because his means had been changing, so we kept following it all the time. So there is a change of thought, that is to focus on the object we want to protect, that is, the right part of my picture, the assets, the people in our work environment, our process, which are actually the core of our protection. Once we focus our attention on the objects we want to protect, we really have the chance to remain unchanged. This gives us the opportunity for AI applications, and some basic concepts are that we have to use a different way of modeling from a monitoring point of view, and then find out if there is an unusual condition, and this is a simple example. If we know about data protection from data protection, privacy protection is a very very important topic. Its similar from the point of view of data protection, that is, we do a lot of different classifications of data, because we know the importance of it at the time of classification, knowing who should use it and when to use it, the concept is back to a topic that we can use AI to do modeling and testing. So actually, to the last point, I mentioned the perception of data leakage. Here is a very, very large AI application space. Look at the safety of the car, as you know, in the traditional taxi platform, the problem of safety is always there, or because of the conflict, or because we have some crimes, on such a big platform, because we have a large number of data and trajectory, at the same time We have the opportunity to make a portrait of everyones driving habits, some of the whole ecosystem of the car, and make a model. On the basis of this model, we can find out the problems in the process of driving in the process of driving. Possible problems can be solved through timely intervention and response. When problems arise, we can send the best response service to the scene at the first time. So this is a AI application scenario from the perspective of physical security. Back to today, we all know that many people are talking about the concept of intelligent driving vehicle and the whole intelligent vehicle networking. Actually, this picture looks rather complicated. I would like to give you a general understanding of it. There is no need to go into the details of it one by one. What is the general concept, that is, from the traditional car to the development of todays car, everyone has a variety of systems, people develop this car into an automatic driving car, further to the car network, meaning between the car and the car, between the car and the road, the environment it can have communication, let our traffic safety. The patency of all aspects has improved to the whole of the intelligent transportation system, which may not only be the vehicle of the smart car, but also the other means of transportation. At each stage, when we introduce such a complex system that may be a better application system for us, we are constantly introducing complex technical factors. With the coming of this factor, it will pose some security challenges to us. We only solve all these problems of security challenges, which are the dark blue or the right box, which is the scheme and the dependant of each technology, and we must make it such a system. In the end, we can make a green frame, that is to say, the guarantee is basic, the security of people and property, and the ecological security, because the car itself can affect the whole environment in an environment. The security of the data is not to be said, from the driver to a variety of vehicles, so from this system to consider the best solution to the security of the best this time is a big expansion of the AI technology. In fact, a core concept gives everyone a concept, especially for students who know safety. First of all, the traditional car itself also attaches importance to safety. It attaches great importance to physical safety, which is the design of functional safety. Today, when we introduce intelligent driving, there is a higher demand, because what, because when we detect a situation, we used to respond to a driver, a person, and today it may respond to another intelligent machine, so this equation is completely changed. Second, the main ingredient here is to use machine learning to learn all kinds of models, and then do a Inference thing to make many decisions. We have to be aware of todays machine learning, no matter what kind of algorithm it has, and what it means, and what it means, it introduces another topic, that is, the third, that is, we actually have to consider it from a security point of view, and the face of the attack is actually much more, that is, the traditional variety. Loopholes, operation errors, but increase the application machine learning, have large data as input to build the model, model learning process also has a concept, that is, the concept of a hostile machine learning, it wants to attack you, it can create a variety of scenes, making this difficult. What it means is that we have a core principle when we solve this problem. This principle is to think of a systematic thinking. Here we have to consider a larger environment, and in the process of doing it, the design must take into account what the case is. In the event of a situation, we must have a means to ensure safety, which is the core concept. I have just talked about security, from the perspective of experience, this is the second, and AI can be of great use. One of these concepts, as you know today, can make use of all kinds of technologies, from voice to image, to various technologies, to many channels to communicate with drivers and passengers. It is difficult to imagine this in a traditional way, and today it means how we can use such a means to do it. A smooth interaction enables us to be more integrated and safer in the driving environment. We also know that we are in China, the world is also the same, China is also very rampant, that is, the whole of our black production, this problem for the whole driving safety environment is also a great impact, here is a problem we have to face is the problem of man and car. There are many reasons why the human car does not match. The most important thing is that there is a black product, and the information can be stolen through the illegal way. This information may be used by criminal gangs. It can make various means of changing cards and numbers. Then there are some drivers who turn themselves into such a small bad ecosystem for personal gain. In view of this ecology, we can take a series of systematic measures by using AI technology. Here is an interactive process based on a variety of AI intelligence, which eliminates a lot of such a scene, including the technology involved in face recognition, the voice technology, the whole problem of the security of a registration phase, as well as the safety of the driving link, such as the requirement. The driver and the passengers take photos together, and then we can go through the automatic technology to do a verification. This simply gives you another example. When we use the strategy and AI, we have seen a good effect. This effect is actually a means of making the blackout they use. They also saw this problem, in fact, they had been blocked by a dialogue. The last point is the third aspect of the application of AI, that is, the improvement of the overall efficiency is that we do things better. The first scene, some students may have experience, is a carpool concept, can imagine our same driver, a car in a relatively fixed period of time if we are from a classmate, one person, a passenger can get to three passengers, can imagine our efficiency at this time. Do not consider the others, at least three times that is a core concept, which is to do this, and it is so complex that the complexity of its algorithm is very large, because the whole car is walking, the person is going, but the destination we have to do a match in a very short time, so that this scene is the three people to ride. Take the same car. Another concept, everyone has taken a taxi, and the other will have such experience, that is, often in the car when it is difficult to find the place, this is also through the AI data learning, in fact, we have continuous feedback, the whole system can be found in the whole system, in different places, different. Location, is POI, is the place that everyone is interested in, take this as the car, call the cars goal, so the experience of everyone will be of great help, this is to accumulate through the AI technology, it is a dynamic learning process. Finally, I would like to make a point about efficiency. This is our work or more advanced research. First, in fact, we have more than 20 cities in China, which have already been deployed and have been put into use. This is the concept of intelligent traffic and traffic lights. We can build a very accurate model of traffic flow through the data collected on the platform, and then we can learn by machines again. The best strategy is to let us control the traffic lights in the entire area to help coordinate. In this way, we can see that in the 20 cities experiment, we can see the mitigation of 10%-20% congestion, this is a very substantial figure. At the same time, using the same concept, we can let the car see and hear no AI technology, it may not hear, see the scene, which not only affects the peoples experience, but also affects the safety of everyones driving. So this is what I show to you on the right, that is to say, through such technology, we can hear and see scenes we cant see at all. We really want to use technology to change traffic. This is indeed a great challenge, which is not something that one can do. At the same time, to ensure security, indeed we have done security for so many years have been deeply experienced, that is, the problem of security in todays extremely interconnected years can not be an island security. This means that we can solve this problem by means of an ecological concept, an ecological approach and an ecological approach, and return to a win-win concept. Today, look at the drops. In fact, we probably know that we have an investment in the world, except that we have a lot of travel related. At the same time, we have a business in Brazil and Mexico. One of the interesting things is that there is a big newspaper in Brazil that reports the whole smart traffic. After the impact on urban traffic, many cities with heavy congestion have come to find out how we can cooperate. This is what we see from the perspective of win-win situation to solve our biggest challenge. This is what we have seen. We are in North Americas first research room, in North America, where we are in charge. We have more than 100 students now, not only to communicate with you, but to let you know that we have done a lot in China, in fact we havent done anything yet. It also attracts a lot of talents. I hope you can work together and have a chance to go there. Thank you. (end) pay attention to NetEase intelligent public address (smartman163), for you to interpret big events, new ideas and new applications in AI field. Source: NetEase intelligent editor: Guo Yin Shuang _NBJS5731 This is what we have seen. We are in North Americas first research room, in North America, where we are in charge. We have more than 100 students now, not only to communicate with you, but to let you know that we have done a lot in China, in fact we havent done anything yet. It also attracts a lot of talents. I hope you can work together and have a chance to go there. Thank you. (finish) Pay attention to NetEase intelligent public address (smartman163), to interpret the big events, new ideas and new applications of large companies in AI field.