Fusheng AI, the hottest cheetah, will greatly impr

Cheetah Fu Sheng: AI will greatly improve social efficiency

original title: Cheetah Fu Sheng: AI will greatly improve social efficiency. On December 28, Fu Sheng, chairman and CEO of cheetah mobile, attended the people's Conference on artificial intelligence partners, And delivered a speech on the theme of "robots help upgrade China's service industry"

in his speech, Fu Sheng said that the development of artificial intelligence has greatly improved social efficiency, and turned the different professional knowledge in each industry into a unified underlying technology, and the boundaries of the industry are constantly disappearing

Fu Sheng believes that compared with technical ability, the landing application ability is the biggest opportunity for Chinese AI entrepreneurs. "Although we are doing artificial intelligence technology, we must find a place to improve technology with the needs of the market and users."

the service reception robot leopard Secretary began mass production in December. However, Fu Sheng believes that mass production has only completed the first step, and the most important step is to really land on various scenes, so that users can really use it, and achieve better standardized services while reducing employment costs. At present, from intelligent government affairs, intelligent navigation to intelligent conference affairs, cheetah mobile robot intelligent solutions have been implemented in multiple scenarios, and bloomed all over the country

at the same time, at the AI Partner Conference, the people announced the establishment of AI research institute, and Cheetah mobile and more than a dozen other AI enterprises reached strategic cooperation. In the future, cheetah mobile will work with industry partners to jointly promote the upgrading of China's service industry

the following is the full text of the speech:

it is a great honor to participate in the AI partner conference hosted by the people today. Although cheetah mobile is a very young company, it has long been a global company. We were founded in 2010, when China's mobile Internet industry was in a sea of blood competition. I realized that the competition in China was too fierce, so we looked around the world and found that at that time, the global mobile Internet was also booming

when I first went to the United States in 2012, I was surprised to find that the apps on the U.S. tool and security software rankings are difficult to use, and there is a gap between them and what we do ourselves. I realized that in the field of mobile application development, China has begun to take the lead unknowingly

the boom of mobile Internet has receded, and AI will be a new technological revolution

at that time, I was wondering whether there was a chance to use mobile Internet to become a global mobile Internet company. Therefore, unlike many Chinese companies, we give priority to the development of overseas strategies. From 2012 to 2014, it took us two years to increase the number of mobile Internet users from zero to more than 400million per month. Today, cheetah Mobile has nearly 600million monthly active users, 70% of which are from overseas, accounting for more than half of overseas revenue, and 20% of our users are from European and American countries, which is the characteristics of our company

we began to invest in American high-tech funds and Israeli funds very early. When I was listed on the New York Stock Exchange in 2014, I gave myself a proposition: if the wave of mobile Internet is over, what will be the next wave

I didn't come from a technical background, but I was deeply shocked by alphago. I realized that AI may be a technological revolution. I think the significance of this technological revolution has two points:

first, efficiency has been greatly improved

second, the different specialized knowledge in each industry in the past has been turned into a unified underlying technology

AI has gradually eliminated the knowledge barriers and skill boundaries of all walks of life, making it possible for a company to not only have the ability of visual recognition and speech recognition, but also to show its skills in specific industries such as driverless driving, which is its charm. Moreover, it brings all people's wisdom together. When the underlying algorithm and chip development become a unified underlying architecture, the burst force of artificial intelligence may be very strong

moreover, I also realize that although AI is a technological revolution, it is not something that a large company can monopolize. On the contrary, it makes the accumulation of many technologies in the past essentially meaningless. For example, after the emergence of artificial intelligence, the accumulation of face recognition using pattern algorithm will be instantly replaced by the field of utilization

therefore, companies that are not too big like us also have the opportunity to overtake in corners. We thought we must all in AI at that time. Since I was a child, I also like to play with machinery. Later, I also thought that maybe robots are a better entry point

macro policies promote the development of artificial intelligence industry

we are also very lucky to see that the country has issued a variety of policies, which greatly support artificial intelligence

from a higher level, my conclusion is that the first 40 years of China's reform and opening up have been 40 years when basic necessities of life have been met. In essence, because we have a huge gap in the basic level of food, clothing, housing and transportation, eating, drinking and playing, so as long as we operated extensively in the first 40 years, we could achieve great volume growth

although I am not engaged in economics, I once discussed the "Income Trap in medium developed countries" with an entrepreneur friend. When the per capita GDP of many countries reaches 10000 US dollars, there is a bottleneck. We analyze from the perspective of enterprises why this is the case and why other countries have crossed this barrier

later, we found that if the per capita GDP reached 10000 US dollars, the basic needs of food, clothing, housing, transportation, food, drink and entertainment might be met. At this time, the only chance to cross this trap may be the breakthrough of technology, find the landing point of new technology, use new technology to improve the efficiency of the whole society and enterprises, and create new demand with new technology

although economy is a very macro thing, it must be composed of micro

from a macro perspective, the state has introduced such regulations and policies at this time to support the development of the artificial intelligence industry. Entrepreneurs like us feel very lucky and feel that this is a great help. Because there is a big gap from the emergence of a technology to the landing of products and finally the formation of industrialization. Sometimes, the crossing of this gap depends on the market, but the support and guidance of the state, as well as supporting industries, are also very important, which may shorten the generation gap of technology

from a micro perspective, if we find such a wave of technology, how can we really turn such technology into an industry. AI has very high requirements for technology, but after I really entered the field of AI, I found that it is actually the same as Internet. I personally believe that the deep learning we define today should be implemented in specific scenarios

defining the scene and landing well are the biggest opportunities for AI entrepreneurs.

everyone knows that mobile payment is one of the new "four inventions" that Chinese people are proud of. Our advantage today is to truly make the ease of use of the payment scenario the first in the world. The ability to implement applications has been the biggest opportunity for Chinese Internet companies, and even the entire Chinese entrepreneurs. So our own team said that although we are doing artificial intelligence technology, we cannot simply talk about technology. We must find a place to fall. We must use the market and users to drive the improvement of our technology, and we should have a faster iteration speed

our technology is accessed by Xiaoai classmate of Xiaomi. Now 80% of the voice to text work of Xiaoai classmate is handled by us. At the beginning, it was connected to 10 voice to text engines. Our technology ranked tenth on the first day of access. We became the first in a month. In fact, the core is to rely on rapid iteration and scenario based iteration

today, the whole Internet in China is so developed. In fact, the essence of Internet is still a service industry. We can arm the idea of Internet with AI and turn it into a landing product. There is a problem in the industry. It is a hard and outstanding work to truly replace people in some vertical scenes. At that time, we chose to be robots in the service industry. And I also put forward a word called "really useful" robot. The first is to define the scene clearly. If we emphasize today that a robot can also be used, at least I can't do it. For example, our service reception robot leopard secretary, as long as we do the reception and guidance in this venue, we can achieve the replacement of people to some extent

using Internet technology to promote the upgrading of China's service industry

was recently swiped by an article in the circle of friends, which said that our birth population has been declining this year, down to more than 10 million, and China may enter aging earlier

in addition, the post-90s generation is unwilling to work in the service industry, and the turnover rate in the service industry is very high. Today, fewer and fewer migrant workers enter the city and fewer cheap labor to fill the gap in urban life. We may only have the opportunity to complete the change of social structure by means of technology

at the same time, it is difficult to maintain a high level of service. But the machine can do it. It may be stupid at first, but all users' problems will become scene data. The machine updates directly in the background in an interconnected way. The more complete its knowledge base is, the better it will perform in a vertical scene

therefore, I think AI is a premise for us to talk about robots today. Robots must not only be mechanical manufacturing, but also mechanical manufacturing is only a part; It's not just about interconnection, because without the AI interface, the device doesn't have the ability to perceive. It won't know who is coming and what he said

today's robot is essentially a support system based on AI, coupled with the coordination ability of hardware, the continuous customization ability of software, and the real-time update of big data services in the background

we used to say that the efficiency of Internet is high because all user behaviors on the Internet are data. Your click, stay and search have become the basis of Internet decision-making. We say that offline efficiency is not high enough because everyone comes, and you don't know how many people come. Today, thanks to AI, the camera really starts to know people, and the microphone can really turn the user's query into search. At this time, it can actually become as efficient as online and more scene oriented

so why are Internet companies laying out offline and taking the road of industrial interconnection today. In addition to the fact that the dividend competition of the Internet has reached a certain degree, there is also the achievement of technical means. AI can really make the efficiency of online and offline as high as that of offline, and offline is more scene oriented

the robot industry is still in its early stage

cheetah Mobile has invested in the establishment of artificial intelligence company Orion star, which focuses on the underlying capabilities of AI and the manufacturing of robot equipment, but today's robot industry, or artificial intelligence industry, is in its early stage. Just now, I told Mr. Zhao that he said you made your own indoor navigation? I said yes. Originally, we wanted to buy a robot chassis and install it on it. We found that the chassis might cost 100000 yuan, because the technology is too early, and the algorithm