1. Technical standard: BS EN 10202, DIN EN 10203, GB/T2520 , JIS G3303 Printed Tinplate Sheet,Packing Printed Tinplate Sheet,Tin Free Steel Sheet,Tinplate Coil Color Lacquered Jiangsu Guolian New Material Co., Ltd. , https://www.cntinplate.com
This is also the story that is happening in China. From Silicon Valley to Beijing, artificial intelligence is a heat wave. This technique, which set off two climaxes in the 1950s and 1980s, seems to have really reached the critical point of industrialization. Kevin Kelly, a thinker and prophet in the Internet age, claims that artificial intelligence is the next technology that has subverted human society for the next 20 years. Its power is comparable to electricity and the Internet.
Artificial Intelligence, abbreviated as AI. It is a new technical science that researches and develops theories, methods, techniques, and applications for simulating, extending, and extending human intelligence. Research in this area includes robotics (26.660, 0.12, 0.45%), speech recognition, image recognition, natural language processing, and expert systems.
Artificial intelligence quickly entered the public's field of vision, which originated in March this year, Google Go AI artificial program AlphaGo beat the world champion Li Shishi. The game of man-machine Go is just a show of popularizing artificial intelligence. Behind it is the scale of the artificial intelligence industry market. The BBC predicts that the global artificial intelligence market will reach 119 billion yuan in 2020.
At present, this data is far less than the scale of China Mobile Internet in 2015. However, the significance of artificial intelligence is not only this, the development of artificial intelligence technology will also drive the upgrade iteration of cloud services, big data analysis, mobile Internet and Internet of Things industry. It will even transcend the mobile Internet and completely change the way people live and work.
The IT Internet company closest to artificial intelligence is keenly aware of this opportunity. Since 2011, the mature technology of deep learning algorithms, computing resources and big data industry has made a leap forward. Domestic and foreign companies including Microsoft, IBM, Google, Facebook and BAT have begun to deeply deploy artificial intelligence, trying to grasp The feng shui has become the giant of the next industrial revolution.
IT and Internet giants who have deployed artificial intelligence are best qualified to be the beneficiaries of this revolution. But this is a long endurance game. In addition to the technical layout, industrial layout and strategic vision are also the key factors determining success or failure. Short-sighted people will be out at any time.
BAT conservative layout
Among all industries in China, the Internet industry represented by BAT has been the earliest in terms of artificial intelligence research and commercialization exploration, and it has also seen the furthest. Among them, Baidu, which started from technology, is at the forefront.
In October 2012, Baidu Chairman and CEO Li Yanhong participated in an internal deep learning-based speech recognition product research conference. At that time, Yu Kai, the leader of the product, recalled that it was Li Yanhong who first learned deep learning. He was very surprised and wrote a letter to the whole company, so that all product managers should understand the development of artificial intelligence technology.
In December of that year, Li Yanhong began to discuss with Yu Kai the feasibility of establishing a deep learning institute. In July of the following year, the institute was established, Li Yanhong was the dean, and Yu Kai was the executive vice president. This is the first artificial intelligence research institute in China.
Li Yanhong quickly put the relevant technology into the core business of search. Baidu's internal statistics in 2014 show that the application of deep learning technology has increased Baidu's and competitor's Diff (different, internally called Diff) indicators several times.
But the research of artificial intelligence, especially basic scientific research, is a long and lonely process. Since then, due to financial data and competitive pressures, Baidu's decision makers have paid more attention to the immediate layout and resources.
"In the later period, too long-term projects, or relatively innovative projects, Baidu headquarters does not support much. Baidu i station project, Baidu Express search projects are gone." A former Baidu artificial intelligence research that does not want to be named Post evaluation. A current Baidu related person told the Caijing reporter that Baidu had indeed eliminated many branches that have been verified to have no commercial prospects in the past few years, but in the last two or three years, Baidu has significantly increased its investment in artificial intelligence. Includes long-term projects such as driverless cars.
On June 8, at the 2016 Baidu Alliance Summit, Li Yanhong called Baidu driverless car "a computer with wheels." He broadcast live video of Baidu unmanned vehicle road test. Baidu unmanned vehicle can accelerate, parallel and overtake like normal vehicles. He also said that within three to five years, driverless driving can become a reality.
Overall, Baidu is still the first company in BAT to complete the integration of artificial intelligence technology systems. At present, technologies such as Baidu Research Institute, Baidu Big Data, Baidu Voice and Baidu Image have all been classified into artificial intelligence technology systems. Li Yanhong repeatedly stressed to the outside world that the future development of Baidu will rely heavily on artificial intelligence.
An artificial intelligence expert at the IBM Research Institute told Caijing that Baidu was the only Chinese company to be included in the competition list.
Silicon Valley is particularly concerned about the progress of the "Baidu Brain." Baidu brain is the core of Baidu in the field of artificial intelligence. Many of the artificial intelligence products previously released by Baidu, such as driverless and intelligent search, are based on the capabilities of Baidu's brain.
Bai Jin, senior vice president and general manager of the autonomous driving division, said that Baidu’s brain has the ability to see, hear, speak and predict, plan decisions and control actions. In terms of data, Baidu has trillions of web pages, mobile and behavioral data for analysis. In terms of deep learning, Baidu's trillion parameters rank first in the world.
In 2015, Baidu's R&D investment exceeded 10 billion yuan. Accounted for 15% of Baidu's total revenue of 66.382 billion yuan in 2015.
Baidu is planning to commercialize Baidu's brain in the fields of finance, automobiles and medical care. Li Yanhong said that the "blowout" innovation of artificial intelligence will push the Internet into the third act and will reconstruct the traditional industry. For example, “Artificial Intelligence + Finance†can quickly upgrade credit information and realize “second shift†loans.
The layout of Alibaba and Tencent is more restrained. Or, they are more representative of the common practice of Chinese companies, starting from the business drive, and gradually increase the gear.
Since 2011, Ali has been laying out hundreds of companies in the field of Internet medical, investment acquisition and strategic cooperation. A lot of layout has been done around hospitals, medical insurance, and medicine. The latest statistics show that more than 400 large and medium-sized hospitals across the country have joined Ali's “Future Hospital†program, covering 90% of the country's provinces. Alibaba Cloud's chief scientist Dr. Wan Wanli told the Caijing reporter that Ali's layout in the health care field depends on Ali's technological breakthrough in the field of artificial intelligence.
Ali's vision is that in the future, in Ali's hospitals around the country's remote mountain villages, the patients will not leave the house, just take a CT and complete the expert-level diagnosis and treatment process through remote technology. This kind of diagnosis and treatment depends on Alibaba Cloud's artificial intelligence medical system. As learning data continues to increase, machines become smarter and become a “medical expert who never retiresâ€.
Many experts who were interviewed by Caijing reporters commented that Ali's technology is not complicated. Medical diagnosis is based on experience-based expert labor. It is an experience that the machine is good at, and the technology can mature in three years.
Wan Wanli told the Caijing reporter that in order to achieve this goal, in addition to the technology platform, the entire medical system needs to be opened up, and the government and the society should jointly promote the opening of equipment and resources. Once opened, aggregation on an artificial intelligence service platform will lead to more application service scenarios, thus achieving "business and technology drive each other."
Ali is the most complete and sound data company in all Chinese companies. Yang Yang, founder and CEO of iPIN, believes that Ali's MaxComputer data channel is a very healthy data artery that can integrate all of Ali's data resources very efficiently.
In addition, Alibaba Cloud is also the closest cloud computing platform in the world to AWS (Amazon Cloud Service). The data ecosystem is an important foundation for artificial intelligence. Therefore, in this future layout, Alibaba Cloud's dominant position is clear.
Ali's artificial intelligence research is scattered among its various business branches and needs to be integrated. Wan Wanli said that Ali hopes to make a breakthrough in algorithmic capabilities, which requires a centralized technical organization to integrate all relevant technical resources of Ali.
Tencent and Ali are similar. Tencent's layout on artificial intelligence is based on IM and SNS services. For example, voice recognition is mainly in the WeChat department, picture recognition is mainly in QQ, payment and financial services are implanted face recognition, and the search department is concerned about natural language recognition.
Some of these technologies have been commercialized within Tencent. SNG (Social Network Business Group)'s excellent map team focused on the field of image recognition, launched a yellow map recognition function, and provided technical support for Tencent's internal products such as the image optimization tool "Daily P". WXG (WeChat Group) is an extension of the human-computer interaction field, and it also carries out the original image and speech recognition. WXG launched the intelligent robot "small micro", users can communicate with them in natural language to solve the shortcomings of the previous voice assistant intelligent mechanical response. For the future, engineers hope that "small micro" will become a "connector" - in tandem with WeChat payment, access to WeChat public number and various life services in the wallet, to create a complete WeChat O2O closed-loop ecosystem.
Yao Xing, senior vice president of Tencent, said in an interview with Caijing that Tencent is paying more and more attention to technology development in the field of artificial intelligence. This includes two paths: First, integrating Tencent's own technical resources, forming a system and focus; It is to accelerate the pace of acquisition and cooperation of outstanding companies.
Tencent is involved in early investments in multiple artificial intelligence projects. Tencent Investment M&A has been looking for the necessary targets for the company and serving Tencent's overall strategy. Tencent recently led the Silicon Valley venture capital firm Felicis Ventures to lead the $10 million A round of the artificial intelligence startup Diffbot. Through artificial intelligence technology, the company allows “machines†to capture key content of web pages and output structured data that software can directly identify.
BAT's artificial intelligence technology development is driven by business demand from the first day. They start from the business, buy and even imitate other people's things. This approach avoids aimless research and unnecessary failure, but it cannot guarantee the next one. The round of artificial intelligence platform won the battle.
This year, Ali and Tencent have the idea of ​​forming an artificial intelligence research institute. Yao Xing told the Caijing reporter that soon Tencent Artificial Intelligence Research Institute will be established.
International giants go deep into no-man's land
If the artificial intelligence layout of BAT is in the state of benchmarking and catching up, then the American giant companies represented by IBM, Microsoft, Google, and Facebook have begun to penetrate into the technology no-man's land.
These companies have different technologies and businesses, and they target different users, but their goals are the same: to make artificial intelligence machines bigger, stronger, and better.
IBM and Microsoft may not have Google, Facebook looks so cool, but in the field of artificial intelligence has a deep technical background, IBM has even begun to use artificial intelligence to make money.
IBM's artificial intelligence research dates back to 1997 when "Deep Blue" defeated the then chess world champion Kasparov. In 2011, Watson, who represents IBM's most advanced technology in cognitive computing, became famous on a TV show and is considered a milestone in the history of artificial intelligence.
Today, on behalf of IBM's highest level of expertise in artificial intelligence, is the ever-evolving Watson system, and the already-productionable human brain analog chip SyNAPSE (very large-scale synaptic computer chip).
Watson is a supercomputer originally composed of 90 IBM Power 7 servers in parallel. Compared with Google and Microsoft's artificial intelligence, it began to simulate human neurons from the hardware chip architecture, based on IBM's "DeepQA" technology development. In early January 2014, IBM announced the formation of the "Watson Group" to further develop, commercialize and enhance "Watson" and other cognitive technologies, in addition to investing $1 billion in other related projects.
Watson has started making money for IBM. The French Agricultural Credit Bank predicts that the income generated by the Watson system will account for more than 12% of IBM's total revenue in 2018. Watson has been deployed on Softlayer, the cloud computing infrastructure business acquired by IBM last year, and has become a weapon for IBM to compete with large technology companies such as Amazon, Google and Microsoft in the field of cloud computing.
Another representative product is the human brain analog chip SyNAPSE released by IBM in 2014. The chip can mimic the working mode of the human brain, low power consumption, and is far superior to the traditional computing architecture in cognitive computing. Unlike other chip companies' paper plans, this chip has reached mass production requirements.
IBM has summarized its technical and commercial strength as a "cognitive computing system." Guo Jijun, vice president and general manager of IBM Greater China, told the Caijing reporter that IBM is promoting cognitive computing systems with the goal of accumulating IBM in all areas of artificial intelligence, big data, deep learning, and pattern recognition. Apply to various industries to help customers in all walks of life improve efficiency and solve the real challenges they face.
Microsoft's artificial intelligence technology has been researched for more than 25 years. In 1991, Microsoft established the Institute, the first five research groups, the research direction is human-computer interaction, natural language processing and machine learning, speech recognition and speech synthesis, computer vision. These are just a few of the most important branches of artificial intelligence today.
Microsoft's artificial intelligence research direction is much broader. Microsoft Research has more than 1,000 scientists, and the technical layout in many fields including deep learning is at the top of the world.
Microsoft's latest deep learning system reduced the computer vision system error rate to 3.57% in the 2015 ImageNet Computer Vision Recognition Challenge, which is the first time artificial intelligence has identified an image error rate compared to 5.1% of human eye recognition. Beyond the human level. These machines are powered by Microsoft's Azure cloud service.
Microsoft not only applies artificial intelligence technology to core businesses such as Windows and Azure, but also builds an open platform to open up many years of technology accumulation to the industry. Its goal is to create an artificial intelligence ecosystem.
It travels the furthest in the no-man's land and is the deepest hidden in the real business world.
Compared with the predecessors, the younger Google’s work in the field of artificial intelligence makes the outside world more understandable and more exciting. On the one hand, Google tirelessly does the accumulation of underlying artificial intelligence technology, develops more advanced deep learning algorithms, and enhances graphics recognition and speech recognition capabilities. On the other hand, the company has arranged for smart home, autonomous driving, robots (acquisition of 8 robot companies in 2013), the former provides basic technical support for the latter, and the latter provides data and feedback for the former.
It is worth mentioning that Google's technology accumulation in the field of driverless cars has far surpassed that of traditional car manufacturers and other Internet companies.
Younger Facebook sees artificial intelligence as one of the three major directions for the future. Facebook naturally has massive social data on a global scale, but it still spares no effort in basic science research. In 2013, Facebook established Facebook AI Research (FAIR) in California. Dr. Tian Yuandong, a Ph.D. in robotics at Carnegie Mellon University and a researcher at Facebook's artificial intelligence group, said that FAIR's research direction is free and relaxed. The computing resources (such as GPU) required for research are relatively abundant, and there is no recent product pressure. It can be difficult to focus on the long-term. The study of the essence. He said that such an academic atmosphere is extremely rare in major companies.
If the aforementioned giants are all based on artificial intelligence technology, combined with cloud computing to give technology more potential, then Amazon's path is just the opposite. Amazon is the world's largest cloud service provider, and its cloud service revenue exceeds the sum of all rivals such as Microsoft, IBM, Google, and Salesforce. However, most of the artificial intelligence technologies currently released by Amazon are focused on the deep learning field that enhances the shopping experience.
Artificial intelligence technology has two major elements: the core technology platform and the data loop. It is not enough to have technology. It requires a combination of business and data to create a good technology. For the acquisition of circular data, the giants have spared no effort.
For example, Microsoft, the most enthusiastic open source, last year, Microsoft released the "Oxford Plan" (now renamed "Microsoft Cognitive Service"), which is a smart API (application programming interface) based on the Microsoft cloud platform, covering the top five The direction of artificial intelligence technology, including computer vision, voice, language, knowledge, search five categories of API. How-Old.net, which was popular last summer, is an application that was quickly developed with the platform, with a total of more than 20 lines of code.
Similar tools are available from Microsoft, such as the Deep Learning Toolkit (CNTK) and the Microsoft Distributed Machine Learning Toolkit (DMTK), led by Microsoft Research Asia.
These are quite practical for startups and SMEs. They don't have to learn a little bit from the underlying technology, they can be called directly on a small cluster or on a cloud service.
For the giants, the algorithm is no longer a barrier to competition, data and user habits are the hills. A large number of start-ups will use open source to do a lot of vertical business, including massive trial and error and verification, and will eventually return to open source, which is what the giants expect.
IBM, which does B2B business, has higher requirements for data professionalism. It can't rely on the interaction of search engines and a large number of applications to train the Watson system. Therefore, through deep cooperation and mergers and acquisitions to obtain professional data.
In the medical field, for example, IBM collaborates with several world-class hospitals to deploy Watson's intelligent systems to hospitals, and to help them develop, observe, and adjust by analyzing the hospital's medical records, expert treatment experience, and existing academic research. Treatment options for cancer patients. In the process, Watson also has data accumulation in this field.
In April 2015, IBM acquired Explorys, an analysis company that can view 50 million US patient medical records. There are quite a few similar acquisitions of IBM, and the shots are quite generous.
Watson has already supported initial diagnosis of cancers such as breast cancer, lung cancer, colorectal cancer, and skin cancer. In the field of skin cancer, in a study of 3,000 dermascopic images, Watson identified skin cancer with an accuracy of more than 95%. The accuracy of human recognition of skin cancer is only 84%.
An artificial intelligence industry in China is ridiculous. What the international giants really value in the field of artificial intelligence are their technologies that you can't see, are not open source, and are ambiguous in international conferences. "Those are ammunition that can subvert the future."
Fill the fault
The industrial structure of artificial intelligence can be divided into three layers: application layer, technology layer and base layer. The application layer focuses on the combination of artificial intelligence and various industries; the technology layer is algorithm, model and technology development; the base layer is computing power and data resources.
BAT is good at the first floor. In the hands of BAT, it naturally holds the world's largest data resource. But there are severe faults in the second and third layers. China's research level in the field of artificial intelligence has remained at the innovation level of engineering science, physical algorithms and other engineering sciences. There are few talents and resources in the field of basic theoretical research.
Many Chinese and foreign industry players interviewed by Caijing believe that the advantage of BAT lies in the massive data, and the core gap between the international giants is in technology.
Yao Xing, senior vice president of Tencent, told the Caijing reporter that at the beginning of this year, he and Tencent's investment mergers and acquisitions department reached a consensus and began to inspect a large number of American machine learning platform startups. There are not many technology companies in China; the second is to acquire such a company to quickly fill Tencent's shortcomings in the algorithm field.
Yao Xing analyzed the Caijing reporter and provided 100,000 samples to the machine. The excellent algorithm platform may only take a few hours, and the slow speed may take several days.
For overseas acquisitions, Sogou CEO Wang Xiaochuan is even more straightforward: "There are few domestic companies that are suitable for acquisition, because the roots are broken, and the source of technology and basic research is abroad, and it is necessary to look abroad."
In the fast-paced Internet world, even the Internet giants will miss the opportunity to practice their own secrets. The best way is to have data and technology-owned companies that complement each other through various alliances and quickly seize the market.
In November 2014, Ant Financial announced that it had cooperated with the technology strategy and used the latter's face recognition technology Face++ software to confirm the identity of the user who opened the online bank account, namely “face paymentâ€.
Face++ has won the world number one in multiple benchmarks for face detection. In 2013, it achieved an accuracy rate of 97.27% on Internet news photos that were extremely difficult to identify. This indicator is higher than the Facebook team. After three years, this accuracy rate has increased to 99.5%.
Face recognition requires processing a large amount of data from the face, including data analysis of structure, facial features, and muscles. Alibaba Cloud injects its own data and analysis capabilities into this collaboration.
"Every problem that is spent on solving money is not a problem. Ali can do these things by himself, but the time cost is quite expensive." Wan Wanli told the Caijing reporter, "Ali has 1000 things at the same level to do, can do Good is only a few of them, and the rest is investment + cooperation, which is the most efficient time and capital."
The alliance of technology and data is not limited to BAT. More companies hope to get the future through alliances, and new giants may be born.
Sogou CEO Wang Xiaochuan's idea is to socialize, do more connections, and build the community relationship to express the wisdom in the human brain, thus solving the problem that the current search technology is not accurate enough and practical. After Tencent became a Sogou in 2013, it opened the WeChat public account data and QQ interest tribe to Sogou, and entered data resources for Sogou. In addition, Sogou also knows about 12 million US dollars in strategic investment last November, and fully accessed the content.
Wang Xiaochuan wants Sogou's artificial intelligence machine to constantly learn social data. He told the Caijing reporter, "The next five years of artificial intelligence is not the artificial intelligence itself, but the machine to find people."
Today, the natural openness of technology and data has made the competition among companies “I have you, you have meâ€, and the ultimate winner is a company that can balance technology and data to maximize the benefits of the platform.
Yan Yong, executive vice president of Microsoft Research Asia, believes that horizontal comparisons show that there are gaps in core technologies between Chinese and international leading companies. Foreign countries pay more attention to basic research and technology research and development, and domestic enterprises can tool and commercialize foreign research and development; From a vertical perspective, China's technological accumulation in the field of artificial intelligence has indeed made a leap in recent years. Whether it is the lowest-level computer architecture, intelligent hardware, or upper-layer software applications, there has been qualitative progress.
"As long as it is not too eager to seek success, and consistently investing, China's artificial intelligence industry is worth looking forward to." Yong said.
Extruded foam
In an internal speech, Ma Yun emphasized: "The world is talking about artificial intelligence. When it comes to the cusp, in the face of innovation, there is no second and only the first, innovation is out of date, and you lose."
Not only is BAT anxious, but Ren Zhengfei, the founder of Huawei, spoke at the National Science and Technology Innovation Conference on May 30th. "In the next 20 or 30 years, human society will evolve into an intelligent society. We can't imagine its depth and breadth. If you can't persist in innovation, you will be subverted sooner or later."
Under the open trend, artificial intelligence is also destined not to be a war between giants.
According to statistics from CB Insights, a market research firm, venture capital investment in artificial intelligence increased by 302% in 2014 to reach $309 million.
There are nearly 100 startup companies in the field of artificial intelligence in China, and 65 companies have received investment, totaling 2.91 billion yuan, among which four companies, Defiance Technology, Excellent Choice, Yunzhisheng and SenseTime, have boarded the iResearch Unicorn list.
More startups are just tagged with artificial intelligence. They essentially use international open source platforms, train one or two models with data, and even copy international models. This is actually dangerous. The biggest risk is the serious homogenization of products, especially in mature areas such as face recognition and speech recognition. These companies' products have no breakthrough innovations and simply do not continue to go on or be acquired.
Yao Xing often identifies authentic and artificial artificial intelligence companies for investors. He said that identifying fake artificial intelligence companies has two key points: First, whether the technology used by the company is the latest and most cutting-edge technology, and if not, it is pseudo-artificial intelligence wrapped in artificial intelligence concepts.
Second, is the company's technology and business scalable? If not, it is a commercial company that uses some machine learning algorithms or shallow artificial intelligence technology, not a real artificial intelligence company.
iPIN is a company with text cognitive intelligence technology. Since last year, iPIN has received investment intentions from many investment institutions. Yang Yang, founder and CEO of the company, told Caijing that he has not encountered real An investment institution capable of identifying the level of artificial intelligence technology.
"This is definitely good news for companies that make fake artificial intelligence." Yang Yang said.
The danger is that even some companies that do have the unique technology and business model of artificial intelligence in the initial period are slowly taking shape under the pressure of capital.
Under the pressure of capital, some artificial intelligence startups began to commercialize prematurely, R&D investment gradually decreased, and the personnel structure changed. Sales began to dominate the company, and eventually the technology company became a marketing company, losing the value of being acquired.
Investing in artificial intelligence companies requires professional technical knowledge and long-term investment vision. According to Gartner's “Smart Machine Hype Cycle Diagramâ€, voice recognition is the most industrialized in applications driven by artificial intelligence, and autonomous vehicles and smart consultants are at the highest point of hype, while intelligent robots, natural language processing/generation and virtual personal assistants are at Hill climbing period. These are all subversive technologies that can be widely used in 5 to 10 years. Neuromorphic hardware (such as neuron chips) is a technology that can only be popularized after 10 years, but the technology may have been eliminated before it is matured.
Governments and universities need to be patient and persevering in this round of change. Artificial intelligence involves many disciplines such as computer science, biology, sociology, philosophy, materials science, engineering, etc. The research ability of basic disciplines in Chinese universities cannot be fully utilized. There is a lack of a mechanism for smooth integration of production, study and research in the system. This has led to a serious shortage of Chinese universities in this industrial transformation. From the experience of the United States, it is precisely because of its legal and institutional guarantees that the smooth transition of industry, academia and research has enabled the United States to gain the upper hand in this round of artificial intelligence research.
Some optimistic investors believe that technological development itself is a means of expelling the bubble. "It doesn't have to be very long. In a year or a year and a half, many real situations will be exposed and the bubble will gradually dissipate."
2. Steel type: MR, SPCC
3. Thickness: 0.115 - 0.50 mm
4. Width: 600 - 1050 mm
5. Length: 600-1200mm
Tin Coating: ordinary 2.8 or 2.8g and 5.6g or 5.6g, we can produce according to customer's requests
We can produce as your design
Artificial Intelligence Competition: American Technology Leads China Good at Commercialization
Abstract Artificial intelligence is considered as another technology that has had a subversive impact on human society after power and the Internet. The technology of American companies is far ahead. Chinese companies are good at commercialization and have data advantages. They are struggling to catch up with technology in December 2015, Microsoft...
Artificial intelligence is considered to be another technology that has had a subversive impact on human society after power and the Internet. American companies are far ahead of technology, Chinese companies are good at commercialization, have data advantages, and are struggling to catch up with technology. In December 2015, Dr. Liu Tieyan, Principal Investigator of Microsoft Research Asia, went to Montreal to attend the NIPS Annual Meeting (Annual Conference on Neural Information Processing). Systems), this is the top academic conference in the field of artificial intelligence. But during the meeting, what impressed him the most was not the professional progress of his peers, but a scientist told the audience. In the early years, his doctoral students could not find a job at all, but this year they were robbed.