On November 15th, at the 3rd world internet conference Internet Light Expo in Jiaxing, Zhejiang, more than 300 Chinese and foreign Internet companies showed their strengths. Photo courtesy of vision china
At the just-concluded Third world internet conference, Wang Xiaochuan, CEO of sogou Company, showed a black technology at the Mobile Internet Forum — — During the speech, Chinese-English translation was synchronized through machine speech recognition, and Chinese and English subtitles kept scrolling on the big screen behind him. After that, there are endless discussions about whether artificial intelligence machines can replace simultaneous interpretation.
Wang Xue, a simultaneous interpreter, has been asked this question by several people. This time, even she felt that her career was threatened: "Well, I’m going to lose my job."
"I didn’t believe in big data before, but now I believe it more and more." She believes that except for creative translation of literature, most other translation work can be replaced by artificial intelligence as long as it is based on a database of sufficient magnitude. "For example, Google’s machine translation has basically replaced humans in some fields."
From simultaneous interpreters who have just worked for two years to technology giants Li Yanhong, Yang Yuanqing and Ma Huateng, whether artificial intelligence will replace human work and the development of mobile Internet has become a real change behind the binary code from a joke under the white wall of Wuzhen.
"The next scene is artificial intelligence."
At this year’s Wuzhen world internet conference, the attention of the outside world on the topic of artificial intelligence was triggered by Baidu CEO Li Yanhong. He believes that in the past, the development path obtained by the growth of Internet users was "unreliable", and it is no longer possible to have unicorn-level startups by the enthusiasm of the mobile Internet, and the real rise of the Internet of Things is not far off, and each of us will change because of the arrival of artificial intelligence.
One of the reasons for Li Yanhong’s judgment is that the mobile Internet market has entered a relatively stable stage of development. According to the data of China Internet Network Information Center (CNNIC), as of June 2016, there were 710 million netizens in China, and the Internet penetration rate was 51.7%, up 1.3 percentage points from the end of 2015.
"The dividend period of the mobile Internet has basically ended, and the next scene is artificial intelligence." This is not the first time that Li Yanhong has expressed optimism and expectation for the future of artificial intelligence. At the 2016 Yabuli China Entrepreneur Forum Summer Summit held in August this year and the Baidu World Congress held in September, Li Yanhong clearly expressed his emphasis on artificial intelligence.
Li Yanhong’s prediction is even a bit bold: "In the future, all manufacturing industries will belong to the artificial intelligence industry, or the Internet of Things industry. All goods must be able to be networked, and they must be able to send data back to the cloud. These technologies must be analyzed through artificial intelligence technology, so that it can bring tangible value to our consumers and our users. "
Yang Yuanqing, chairman and CEO of Lenovo Group, and Li Yanhong have the same idea. In his speech entitled "Embracing the New Era of Smart Internet", Yang Yuanqing pointed out that the Internet is entering a new stage of development from PC Internet and mobile Internet, that is, the era of smart Internet.
Yang Yuanqing’s idea is more specific: "Smart devices have become ubiquitous from limited types such as computers and mobile phones, and a steady stream of data has gathered from terminals to the cloud, while artificial intelligence with self-learning ability has become more and more intelligent by constantly reading and analyzing these data."
Based on these judgments, Yang Yuanqing also described a smarter life in the future: home appliances, office equipment, and traveling cars are no longer machines that passively follow instructions to perform calculations, but will be considerate of you and actively serve you. Even the houses and cities where people live will become smarter and more active. For example, windows will adjust themselves according to the outdoor climate, the temperature in the city, lighting, etc. The traffic management system of the city may be connected with the navigation and driving system of the car, thus greatly reducing congestion.
Big bosses verbally affirmed artificial intelligence, and the Internet giants behind them were busy trying more intelligent new technologies.
Lenovo launched MotoZ, which turned the mobile phone into an open platform for hardware. Through unified standards, the mobile phone can be connected to more hardware devices, and then connected to more cloud services. Now it can be connected with JBL speakers and Hasselblad modules; During the period of world internet conference, Baidu unmanned vehicle first carried out open urban road operation in the smart car and smart transportation demonstration zone of Ziye Road, Tongxiang City.
Artificial intelligence is still in the stage of mental retardation.
Although the "black technology" of synchronous bilingual subtitle conversion shocked the scene, Wang Xiaochuan also mentioned a concern in the industry: Will artificial intelligence have a third ebb tide?
According to Zhiyuan Xin’s Report on the Development of Artificial Intelligence Industry in China, since the concept of artificial intelligence was put forward at Dartmouth Conference in 1956, artificial intelligence has developed for 60 years. During the 60 years of development, artificial intelligence has gone through "three ups and three downs": from 1950s to early 1980s, it was still in its birth and basic stage, neuron model was put forward and perceptron was invented; From the 1980s to the early 21st century, artificial intelligence was successfully commercialized but failed to develop by leaps and bounds. Nowadays, with the improvement of computing power brought by Moore’s Law and cloud computing, and the accumulation of massive data brought by the wide application of the Internet and big data, artificial intelligence is expected to achieve large-scale application and will go deep into various fields such as finance, transportation, medical care and industry.
"Before artificial intelligence ebbed twice, we asked a teacher, do you study artificial intelligence? This is to scold him. " Wang Xiaochuan said that compared with the previous two waves of artificial intelligence, the topic of artificial intelligence was once again hotly debated. Artificial intelligence was really used in many industries for the first time, and a lot of money was invested in it. A large number of researchers also engaged in artificial intelligence work after graduation.
However, the current development of artificial intelligence still encounters many obstacles that are difficult to solve. Taking Wang Xiaochuan’s proud speech intelligence as an example, the first problem we face is speech recognition and semantic understanding. Wang Xiaochuan found that many technologies can achieve 95% or even 97% recognition in quiet environment, but the recognition rate drops when there is noise. When two people talk at the same time, human beings can easily distinguish the differences, but it is extremely difficult for machines.
Wang Xiaochuan mentioned that during the testing of a voice intelligent product, they found that the performance of the machine was very poor: the machine asked if it needed a parking space, and the human replied: I don’t have a car. The machine can’t understand this answer.
Hu Yu, president of Flywheel Value of HKUST and president of iFLYTEK Research Institute, also believes that the current artificial intelligence technology is still in the stage of "mental retardation", and most of them can only reach the level of voice assistants. In order to obtain wisdom similar to or equivalent to the human brain and have the ability to learn independently and draw inferences, artificial intelligence also needs to combine deep learning, big data and other technologies to achieve the "ripple effect" — — Through the increase of users, artificial intelligence products can learn and update themselves through real data and experience.
What is the imagination space for the combination of artificial intelligence and mobile Internet?
Ma Huateng, CEO of Tencent, is quite cautious about artificial intelligence that has caused heated discussion. He said that artificial intelligence still needs time, and it is still far from the application of all walks of life.
Ma Huateng said that artificial intelligence is a cloud change and the mobile Internet is a terminal change. Although the growth slope of mobile Internet has slowed down in recent years, its volume is still very large, and there are still many places to innovate. Instead, he feels that there are new opportunities now.
"The current mobile Internet industry is not like three or four years ago. At that time, it was basically a barrier. If it didn’t pass, it would be finished. At present, the growth gradient may slow down, but there are still new opportunities, such as mobike Bicycle ‘ Internet plus Bicycle ’ The mobile terminal suddenly appeared. " Ma Huateng said.
Ma Huateng believes that although the pace of innovation has slowed down since the development of the mobile Internet era, there are still many opportunities for innovation. Artificial intelligence technology is indeed a new outlet, but compared with the terminal innovation in the mobile Internet era, the innovation of artificial intelligence is reflected in the cloud, and there are still many basic technologies to be accumulated. Internet companies still need to combine their own advantages and do more exploration in the field of artificial intelligence.
Yang Yuanqing described the future of artificial intelligence in his eyes in two sentences: intelligent terminals will become smarter and smarter, and the combination of intelligent terminals and cloud will be the future development direction.
In his vision, smart terminals and sensors will be everywhere. Based on the self-learning ability of big data, smart terminals will be extended from today’s very limited types to all devices around people. "Whether it is air conditioners, humidifiers, air purifiers, cameras, cars, machine tools, etc., it will have the ability to calculate storage and network, and it will be modular."
In addition, Yang Yuanqing also believes that the data of these smart terminals will be gathered in the cloud, and a new understanding will be formed through the evolution of algorithms, so as to remember, learn and analyze these data in a more effective way. Through deep learning, various intelligent terminals will become smarter and more judgmental.
"If smart terminals are human senses, then the cloud is the brain. The perfect combination of smart terminals and cloud brains is the future direction of artificial intelligence." Yang Yuanqing said.