How will the AI on the tuyere land?

In March 2016, after AlphaGo defeated the world Go champion Lee Sedol, the concept of "AI (Artificial Intelligence)" was largely focused on the flash, allowing people all over the world to see its influence .

And long before artificial intelligence became a buzzword among the people, Chinese companies were aware of business opportunities. In the past two years, a large number of artificial intelligence start-ups have emerged, giant companies have made frequent moves, and capital has followed enthusiastically.

According to the data of iiMedia Research, through the information collation of 168 non-listed companies with AI-driven technology or business in China, 34.5% of AI startups were established in 2015. In addition, 20.8% were established in 2014, and artificial intelligence established in 2016 accounted for 16.7%. Its analysts believe that 2015 is a veritable year of artificial intelligence startups.

With the rapid development of the industry, the first year of AI has arrived, which is the consensus of the industry. At the conference of the 2017 semi-annual report of iFlytek, a leading artificial intelligence company, Jiang Tao, senior vice president and secretary of the board of directors, said that artificial intelligence is facing an unprecedented window of opportunity, and the entire basic pattern will be determined in the next 3-5 years.

A technological revolution dominated by artificial intelligence is kicking off. The industrial appeal of "AI+" will breed the seeds of change in all walks of life. As the protagonist of promoting industrial development, how should Chinese enterprises seize the opportunity? Bringing technology to application scenarios, digging out real industry pain points and continuing to anchor will be the key for companies to compete for the right to speak.

AI is the next tech dividend?

China's development of artificial intelligence already has the right time, place and people. Xiaomi Lei Jun once said at the national two conferences that 43% of the world's artificial intelligence papers are produced in China, and the emphasis on mathematics education has given China a solid foundation in the field of artificial intelligence.

It can be said that China has a lot of human resources and technological achievements, and its research and development strength should not be underestimated . "In terms of core source technology, we are already in the first echelon in the world, and even leading the development of technology." Du Lan, senior vice president of iFlytek, said in an interview with "CEConline".

According to the McKinsey research report, China will lead the field of artificial intelligence technology, and some Chinese companies have even surpassed well-known global competitors. In a 2017 "MIT Technology Review" list of the "50 Smartest Companies in the World", we can get a glimpse. In this list, which is based on "real innovative technology, and a practical and ambitious business model", there are 9 Chinese companies on the list, including iFLYTEK (No. 6), Tencent (No. 8) Rank), Megvii Technology (11th), DJI (25th), Foxconn (33rd), etc.

These companies are doing well on a global scale, not only with smart technology, but also smart business models. Duran believes that only the rapid application of artificial intelligence technology is the last word of development. In terms of application scenarios, China undoubtedly has unique advantages.

As an institution that focuses on providing hardware R&D services and financial support for AI projects, Shenzhen Chuanggu has invested in many companies with excellent algorithms. enterprise. Its co-founder and CEO Xu Jiabin said in an interview with CEConline: "China has a large number of secondary industry resources, which provide many application opportunities for artificial intelligence. China's large population and large number of enterprises have led us to The application scenarios will be very rich.”

Artificial intelligence will also bring opportunities for the transformation of China’s manufacturing industry. McKinsey & Company President Bao Damin issued a signed report "The Future of Artificial Intelligence in China" and pointed out: With the rapid disappearance of the demographic dividend, China urgently needs to find a new growth engine, and artificial intelligence is an important opportunity for China to accelerate the development of productivity. "Artificial intelligence plays an important role in the transition from 'Made in China' to 'Made in China'," Duran added.

In summary, in China, the development of artificial intelligence has been highly valued. On July 20, 2017, the State Council issued the "New Generation Artificial Intelligence Development Plan" (hereinafter referred to as the "Plan"), which clearly stated that by 2030, the competitiveness of the artificial intelligence industry will reach the international leading level, and the core industry scale will exceed 10,000. 100 million yuan, driving the scale of related industries to exceed 10 trillion yuan. At the policy level, the "Planning" elevates the development of artificial intelligence to a national strategic level.

At the same time, Chinese companies are making efforts in this field. In this list of entrants, in addition to outstanding companies focusing on artificial intelligence, such as HKUST Xunfei, SBI, etc., there are also Internet giants represented by BAT, as well as many well-known companies such as Huawei, JD.com, Didi, etc., among a large number of emerging companies, there are still many unicorn companies. When cross-border enterprises develop technologies, they gradually embed artificial intelligence into various products and technologies of enterprises. Artificial intelligence companies such as iFLYTEK and SPIQI have continuously improved their AI systems through years of data collection, and have made different attempts in various vertical industries, including face recognition, voice recognition, and chatbots.

Standing at the inflection point of the development of the times, China is poised to develop in the field of artificial intelligence. No one doubts that artificial intelligence will be the next technological dividend.

Has artificial intelligence surpassed human intelligence?

What is artificial intelligence? In fact, the term did not come out of nowhere, its theory was first proposed in the 1950s and 1960s. Nils Nilsson, one of the founders of the field, defines artificial intelligence as "the theoretical study of how computer systems can perform tasks that can only be accomplished by human intelligence. For example, visual perception, speech recognition, Decision making, learning, language translation, etc. under uncertain conditions."

The definition in the McKinsey report is that artificial intelligence reads massive "big data" according to a general learning strategy, And discover patterns, connections, and insights. It automatically adjusts to new data without reprogramming.

At that time, artificial intelligence was given the mission of "simulating human intelligence", but due to technological development, its results never met expectations, so with the passage of time, artificial intelligence gradually faded out of the public eye.

In the long period of silence, the technology of artificial intelligence is constantly iterating and upgrading, and people have noticed some of its achievements. However, these isolated cases of technological breakthroughs are still insufficient to support the large-scale commercialization of artificial intelligence.

Until the emergence of the Internet, under the rapid development of the past two decades, a large amount of data that needs to be refined and processed has accumulated, and artificial intelligence technology has made rapid progress. "Machine is like human. To a certain extent, artificial intelligence is 'fed' by big data. Just like a student, if he is not allowed to study, think, and do homework to improve his knowledge, he is It's hard to learn this knowledge point." Duran said.

Accenture's related research report mentioned that the global public cloud computing market has reached nearly 70 billion US dollars in 2015. The further development of cloud computing in the past two years has exceeded the imagination of many people. On the other hand, as more devices become interconnected, the compound annual growth rate of global data volume has reached more than 50%. This means that the data we disclose on the Internet every day will provide artificial intelligence with the "nutrient" for learning.

To what extent is artificial intelligence currently developed? Duran said: "We divide artificial intelligence into three areas: computational intelligence, perceptual intelligence and cognitive intelligence. Computational intelligence refers to enabling machines to 'store and calculate', that is, fast computing and memory storage capabilities. IBM's 1996 Artificial intelligence surpassed humans in computing power in this field after supercomputer 'Deep Blue' beat chess champion Garry Kasparov. Perceptual intelligence is what allows machines to 'hear and speak', referring to vision , hearing, touch and other perception capabilities; artificial intelligence surpasses human beings in some fields. The biggest challenge we face at present is from cognitive intelligence, which means that machines can understand, think, and have the ability to learn independently. In this field, the current scientific community A major breakthrough has not yet been achieved, and the cognitive ability of artificial intelligence is not as good as that of human beings. This is also an area that we need to make breakthroughs urgently.”

In terms of perceived intelligence, there are many intelligent voice companies in China, such as iFlytek , Baidu, Sogou and other companies, their machine speech recognition capabilities are quite outstanding. Among them, iFLYTEK’s speech recognition technology can transcribe speech into text in real time. Under the standard environment, the accuracy rate reaches 98%, which has exceeded the accuracy rate of stenographers.

In cognitive intelligence, Chinese enterprises are constantly exploring and making breakthroughs. It is understood that the “Super Brain Project” being researched by iFLYTEK is an attempt at cognitive intelligence. The plan mainly involves the field of education, and its ongoing 863 major project of the Ministry of Science and Technology - Humanoid Answering Robot Project, plans to allow robots to take the college entrance examination. In the future, machines will be used to assist teachers in marking and marking examination papers, and even marking and marking subjective questions and essay questions.

Du Lan introduced that in terms of core technology, iFLYTEK insists on the innovation path of the source core technology. The company believes that there are three elements for artificial intelligence to change the world in the future: core algorithms, industry big data and industry experts. "First of all, there must be core algorithms. Based on deep neural networks, iFLYTEK has many new algorithms and innovations. Second, there must be industry big data, so that machines can understand and think like humans, they must be fed to it. Massive data. Finally, there are also industry experts. The application and innovation of artificial intelligence in various sub-sectors requires cooperation with industry experts, and top experts formulate standards and gather their knowledge and wisdom for machine learning and training. Machines have the capabilities of the best human beings."

Artificial intelligence scenarios, opportunities and bubbles coexist?

With the continuous development of AI supporting technologies, AI continues to expand more application scenarios. Looking at the AI industry, AI+education, AI+judicial, AI+medical, AI+finance, AI+driving, AI+home, etc. AI is penetrating all walks of life.

According to iResearch, speech and visual recognition technologies currently account for 60% and 12.5% of China's artificial intelligence market, respectively. In China, 71% of all AI-related companies focus on developing applications. The rest of the companies focus on algorithms, with 55% working on computer vision, 13% on natural language processing, and 9% on basic machine learning.

SenseTime, an AI company focusing on computer vision, has captured hundreds of criminals in various places by using the face recognition function of artificial intelligence and the application of big data. It detects, tracks, and recognizes faces in real-time in 1000+ surveillance videos, realizes face control and information mining before, during and after the event, helping the police to solve cases quickly. Face recognition technology has been applied in many fields, including security check, traffic safety, payment and so on.

Similarly, the technology of "vehicle face recognition" has gradually begun to be applied. Through the general details and special details of the vehicle appearance, to obtain vehicle information, such as details such as color, model and appearance defects, it is possible to judge and record specific information, that is, vehicles of the same model and color, through the face recognition technology, can Different owners of the vehicle are determined.

Many AI companies have established business models in different segments and application areas. According to McKinsey, the total value of the artificial intelligence application market in 2025 will reach 127 billion US dollars. In the future, the development of artificial intelligence-related technologies will fully penetrate traditional industries such as finance, medical care, security, retail, and manufacturing, and the market prospect is very broad.

However, the opportunities are plentiful and there are bubbles. Only companies that use AI technology to empower various products and businesses and effectively improve the efficiency of the scene are truly valuable. Otherwise, it seems that the opportunity is ahead, but the bubble is actually hidden.

"The biggest pit of artificial intelligence, I think, is unstoppable. The demand is good, the technology is good, and even the product is good, better than the opponent, but it is not good enough, not good enough to be completely Solve user needs and change the industry structure." Liu Wei, CEO of Baidu Ventures, mentioned.

In this regard, Xu Jiabin suggested that enterprises can look for vertical fields and subdivision scenarios as breakthroughs. One is to find industry pain points and customer needs, and the other is to have typical application scenarios and specific target customers.

"In the future, the AI industry should not be a unified game, but rather each has its own application scenarios and implementation paths, and each has its own development, which is a relationship that confirms and promotes each other." Xu Jiabin said.

Photo/Du Lan, Xu Jiabin