[Ali has also started making AI chips. Will it surpass Intel and Microsoft in the future? "For the past 18 years, Alibaba's business has been so successful that it has overshadowed the brilliance of technology." Zhang Jianfeng, chief technology officer of Alibaba Group, said at the first technology conference held in Hangzhou in March 2017. How successful was Ali's business? On April 6, 2016, Alibaba officially announced that it had become the world's largest retail trading platform. Taobao, Tmall, Juhua cost-effective, AliExpress, Alibaba International Trading Market, 1688, Ali Mama, Ant Financial and other projects just come out any one, are enough to spike many domestic companies. According to Alibaba’s publicly-reported financial report data obtained by Xiao Bian, in the fourth quarter of fiscal year 2017, Alibaba Group's core e-commerce business revenue reached RMB 31.570 billion, an increase of 47% year-on-year; annual revenue increased by 45% year-on-year. 1338.80 billion yuan. This time Zhang Jianfeng stood up and stressed that Alibaba is a technology-driven company that can perfectly integrate technology and business. Then, where does Alibaba's technological brilliance behave? Why did Ali start emphasizing his technicality in recent years? From the X Po in people's cognition to the development of sophisticated AI chips, what did Ali do? Ali also started making AI chips On April 19, Alibaba also stated that Dharma is developing a neural network chip called Ali-NPU. According to Ali’s official description, this new “neural network chip†is mainly used for AI reasoning calculations such as image video analysis and machine learning. At the same time, there is also a very powerful reference number - according to the design, the chip's cost performance will be 40 times the current similar products. Ali did not disclose the specific time for R&D and launch, but stressed that this chip will better realize the use of AI intelligence in business scenarios, improve computing efficiency and reduce costs. In the official description, the description of this mysterious chip is not detailed. It can be confirmed that this chip is a "dedicated architecture chip." This type of chip is different from general-purpose computing chips such as CPUs and GPUs that we are most familiar with. They are more specialized in functions (application scenarios), and because they are more specialized, they will be more thoroughly optimized, and will be better. Performance, energy efficiency performance. Apple, Google, Amazon, Baidu are all developing AI chips Under the hot AI concept, the AI ​​chip is an important battle for each other. Before Ali released the AI ​​chip, Apple, Google, Amazon, and Facebook all revealed plans to develop AI chips. As early as 2016, Google announced that it has produced a chip that can quickly train AI models - tensor processor TPU to fight against Nvidia's round processor GPU, while the old chip CPU giant Intel has acquired AI chip startup Nervana . About this chip of Google, NormJouppi, one of the processor's developers, commented: "This is surely to significantly increase the efficiency of the work." The fact is also true that in the TOPS/Watt (feature per watt) power efficiency test, the performance of the TPU is 30 to 80 times better than the conventional processor; compared to the traditional GPU/CPU computing combination, the TPU The processing speed is also 15 to 30 times faster; the most critical is that due to the use of the TPU, even the number of codes required by the deep neural network is greatly reduced, requiring only 100 to 1,500 lines of code to run smoothly. In February of this year, Amazon also announced plans to develop its own AI chip so that using Alexa-driven products in its ever-expanding Echo product line can do more computing without having to communicate with the cloud. Can increase the speed of operation. According to foreign media reports in March, Apple is planning to introduce a new chip processor in response to increasingly complex artificial intelligence processing tasks on the mobile side. Although Apple has not yet announced the final launch date of the "AppleNeuralEngine" chip, the outside world has learned that it will be integrated into next-generation devices such as the iPhone and iPad to enhance tasks such as facial recognition and speech recognition. Processing power. According to understanding, Apple's AI chip will help the software in battery life, augmented reality (VR), intelligent voice assistant and other aspects to enhance. April 18, according to sources close to Facebook and job advertisements posted on the company’s website, Facebook is forming a team to independently design and manufacture semiconductor products to reduce dependence on chip makers such as Intel and Qualcomm. In addition, several companies in China are conducting research and development of AI chips. The AI ​​industry report released by Tencent previously pointed out that AI chips, as the core of the industry, are also the highest link in terms of technical requirements and added value, and their industrial value and strategic position are far greater than application-level innovation. At this point, the gap between China and the United States is still very large. The report shows that from the perspective of the number of chip companies in the base layer, China has 14 and the United States is 33. China is only 42% of the United States. From the trend of many technology giants, developing their own AI chips and keeping the core of technology firmly in their hands has become a “trendâ€. Is this trend really moving or just rushing? The artificial intelligence market is really big From a global market perspective, according to Tracica, artificial intelligence technology is applicable to almost all industries and will change the existing business models in these industries. In 2016, global artificial intelligence revenue is expected to reach US$643.7 million, and by 2025 it will increase to US$36.8 billion. Judging from the current domestic situation, in the past two to three years, the state has successively introduced many policies on artificial intelligence. From the overall layout of artificial intelligence development goals, encourage the research and development of artificial intelligence technology, build artificial intelligence resource pool, nurture the necessary talents combined with the Internet + model, support the integration of artificial skills and real economy, and promote the development of artificial intelligence on the economy. Figure: Artificial Intelligence related policies as of 2017 In addition, according to the data of the Forward-looking Industry Research Institute's “Market Forecast and Investment Strategic Planning Report of China's Artificial Intelligence Industry in 2018-2023â€, the market size of China’s artificial intelligence increased from 4.86 billion yuan to 13.52 billion yuan in 2014-2017. The growth rate is more than 40%. Among them, the scale of China's artificial intelligence industry reached 9.56 billion yuan in 2016, a year-on-year increase of 37.9%. In 2017, the scale of China's artificial intelligence industry was 13.52 billion yuan, an increase of 41.42% year-on-year! Figure: Size and Forecast of China's Artificial Intelligence Industry in 2014-2017 (Unit: 100 million yuan, %) With the development of artificial intelligence as the main manifestation of competitive advantage at the current stage and the acceleration of China's relevant policies, China's artificial intelligence industry has entered a new stage of development. At present, the basic conditions of the artificial intelligence industry are already in place. With the maturity of deep learning algorithms and the acceleration of data resources, artificial intelligence technology is expected to continue to increase. Artificial intelligence technologies such as machine vision and natural speech processing will usher in new opportunities for development. In the future, with the promotion of computer technology development and the promotion of downstream demand, artificial intelligence companies will gradually expand into downstream demand and promote the further expansion of the market size of artificial intelligence in China. In the future, with the deepening of artificial intelligence, it is forecasted that it will increase at a growth rate of around 10% to 15%, exceed 30 billion yuan by 2022, and reach 40 billion yuan by 2023. Chart: Size and Forecast of China's Artificial Intelligence Industry in 2018-2023 (Unit: 100 million yuan, %) Ali's technological layout began to accelerate From the above data analysis, artificial intelligence does have immeasurable market prospects in the future, and technology leaders are accelerating their layout and consolidating their technical strengths. So what are Alibaba’s technological layouts? Immediately after Ali announced its Ali-NPU chip on the 19th, Ali announced on the 20th that it had wholly acquired the China-based Microsystems Co., Ltd., China's only independent embedded CPU IPCore company. "The acquisition of Zhongtianwei is an important part of Alibaba's chip layout." Alibaba's CTO Zhang Jianfeng said that IPCore is the core of basic chip capabilities, and entering the IPCore field is the basis for China's chips to achieve "independent and controllable". Zhang Jianfeng also stated that in addition to the embedded chip's Zhongtianwei, Alibaba also has a series of investments and deployments in terms of network chips, mobile phone chips, and IOT chips. Up to now, Ali has also invested in five chip companies including the Cambrian, Barefoot Networks, Shen Jian, Kneron, and ASR. At the same time, as an important part of Ali's IOT strategy, Alibaba Dharma has established a high-level professional team to conduct independent research and development of AI chips. In fact, prior to this, Ali participated in the construction of many national key laboratories, including the cooperation with the Chinese Academy of Sciences in 2015 to establish a joint laboratory for quantum computing. In 2017, he collaborated with the National Astronomical Observatory to build the Alibaba Cloud Astronomy Big Data Joint Research Center. Ma has been eager to use it. Ali's technological strength in big data, artificial intelligence and ultra-large-scale distributed computing has made breakthroughs in scientific research. In the past two years, Ali’s technological pace seems to be faster and bolder. In March 2017, Alibaba Group held its first technology conference in Hangzhou, mobilizing more than 20,000 scientists and engineers from around the world to join the "new technology strategy." The meeting revealed that Alibaba is launching a program code-named "NASA", which will "build an independent research and development department for the next 20 years, establish a new mechanism and system, and reserve core technologies for the new economy that serves 2 billion people." Ma Yun mentioned at the meeting that Alibaba’s vision for the next 20 years is to build the world’s fifth-largest economy, serve 2 billion consumers worldwide, create 100 million jobs, and help 10 million enterprises to make profits. Just as NASA's NASA drives tremendous advances in human technology and life, the new economy must be built on new technological infrastructure and built on new technological thinking. Regarding NASA, Ma Yun explained: “An economy serving 2 billion people needs strong technical strength. We will build NASA of Alibaba to take on the responsibility of the future. For machine learning, chips, IoT, operating systems, Biometrics identify these core technologies, we will form a new team, establish new mechanisms and methods, go all out.Our technology has followed the business, is the 'Arsenal mode', but no matter how good the grenades are, they can't make missiles. Alibaba must think about establishing a missile mechanism, set up a new technology research and development system, and focus on the research in the core field. The purpose of these studies is to solve the difficulties in 10 years and 20 years." Ali's NASA program sounds a bit like GoogleX Labs. This is Google’s most mysterious department. They have a list of projects that includes 100 future high-tech ideas such as space elevators. Google hopes to use this laboratory to track these ideas and meet them. Its desire for innovation. Most of these projects are unrealistic. However, Ali is not going to do black technology products such as unmanned vehicles and Google Glass. The establishment of R&D department is only part of the NASA plan. Its purpose is more explicit or more realistic: to reserve all the potential core technologies for all potential business operations of Ali over the next 20 years. It is foreseeable that with the launch of the NASA plan, Alibaba will invest more resources in technology and will invest more in cutting-edge technologies to create a better R&D atmosphere and culture for engineers and scientists. Alibaba will also increase investment in patent applications and other areas, and reserve patents. The NASA plan also marks that Alibaba has officially changed from "ecommerce company" to "technology company." At the 2017 Yunqi Conference in October 2017, Alibaba announced the establishment of the “Dharma House†laboratory to explore the future of human science and technology, and invested RMB 100 billion for research covering basic science and disruptive technology innovation within three years. Foreign media believe that this is an important step for Alibaba to become an international technology giant. For Bodhidharma, Ma Yun said: “The Bodhidharma must also surpass Intel, must surpass Microsoft, and must surpass IBM because we were born in the 21st century.†Bloomberg believes that the Dharma Research Project means Alibaba’s substantial growth in R&D expenditures. This will help the e-commerce giant with a market value of US$470 billion. It will be able to subvert the industry's technology and be able to maintain Tencent and Amazon’s. Synchronize. This is in line with China’s top policy makers’ expectations for China to become a leader in artificial intelligence. Chart: R&D expenditure column chart for Alibaba from 2009 to 2020, in US$ 1 billion According to another comment from Bloomberg, Alibaba’s R&D/sales revenue ratio is currently only half of that of Facebook, and Alibaba’s research investment will have room for growth in the future. Despite the late start and the low starting point, Chinese companies have been among the best in the list of global R&D expenditure growth. According to the 2016 Global Innovation 1000 study, among listed companies, there are one thousand companies that invest the most in R&D. Thirty-three companies came from China and invested a total of 46.8 billion U.S. dollars, a year-on-year increase of 18.6%. Chart: Ratio of Research Investment/Sales Revenue to Global Internet Giants It is understood that at present, Alibaba has more than 20,000 engineers and more than 500 doctors; of the 36 partners, 9 have an engineer background. Prior to this, Alibaba has been investing nearly 10 years in core fundamental technology fields such as ultra-large-scale computing operating systems, mobile operating systems, financial-grade distributed databases, and machine learning platforms. Ali is no longer a "promising" company that only earns women's money and uses Zhang Jianfeng's words: "Today, Alibaba has the ability to be more responsible for making greater contributions to the advancement of human technology and life. In Jin Yong's novels, Bodhidharma represents the highest martial arts institution, and we also hope that Alibaba Dharma will truly become a 'bigger man and benefit the country'. We expect the next disruptive technological innovation like electricity and computers. Was born at Alibaba Dharma." Modern Physics Experiment Series Modern physics experiment related equipment for efficient specialized physics laboratory Modern Physics Experiment Instruments,Optical Instruments,Acousto-Optic Modulator Experimental Device,Optical Spectroscopy Experiment Determinator Yuheng Optics Co., Ltd.(Changchun) , https://www.yuhengcoder.com