Set micro-network launch integrated circuit WeChat public number: 'every day IC', major news instantly released, every day IC, every day set micro-network, plot become a long press! Laoyaoic copy WeChat public number search to add attention.
QUALCOMM and Nokia today announced plans to work together to conduct interoperability testing and OTA field trials based on the 5G New Air Release (NR) Release-15 specification, which is currently being developed by 3GPP, which is designed to drive mobile The ecosystem achieves large-scale rapid verification and commercialization of the 5G new port technology, enabling the 5G-compliant new port network infrastructure and terminals to meet 3GPP standards to support the timely deployment of commercial networks in 2019.
At present, the industry is increasingly focused on 5G mobile applications to meet growing mobile broadband demand and support emerging use cases. In Qualcomm's recent 5G consumer survey, 48% of respondents said they would be willing to 5G smartphone is listed after the purchase.In addition, support 5G as consumers are willing to pay for the next mobile terminal more cost of the primary function.
The two companies will demonstrate a number of 5G new port technology in the test and experiment, efficient implementation of tens of gigabits per second data rate, as low as 1 millisecond delay and significantly improved reliability compared to the current network, and more features. These technologies are critical to meeting the growing consumer connectivity needs and will support emerging consumer mobile broadband experiences such as high-fidelity video streaming, immersive virtual / enhanced reality (VR / AR) and networked cloud computing, as well as autopilot vehicles, Unmanned aerial vehicles and industrial equipment and other use cases to provide both high reliability and low latency of the new service.The two sides also recently jointly issued a white paper "to promote the 5G to become a reality: to meet the 2019 and future strong demand for mobile broadband" , The two sides on the 5G new air port technology, the common vision, that 5G new air technology from 2019 to solve the growing demand for mobile broadband and other needs.
The test and test will be carried out using the Nokia 5G FIRST solution from the Nokia AirScale base station via the 5G new air interface and Qualcomm terminal prototyping machines, both of which will conduct end-to-end OTA testing between base stations and terminals to simulate Real-world scenarios and support for a wide range of 5G new air use cases and deployment scenarios that will cover frequencies below 6GHz (eg 3.5GHz and 4.5GHz) and millimeter-wave (eg 28GHz and 39GHz) to test 5G new air ports in different frequency bands Unified design.
In order to demonstrate the performance and efficiency improvements of the 5G new air port, the test and test will showcase a number of advanced 3GPP 5G new air ports technologies including large-scale multi-input multiple-output (MIMO) antenna technology, beamforming, adaptive independence TDD, OFDM-based scalable waveforms to support larger bandwidth, advanced coding and modulation schemes, and new flexible designs based on low latency slot structures. In addition, the test and test will cover the 5G new airplane running in the millimeter-wave band, Provides robust and continuous mobile broadband communications through advanced 5G new airplane antenna technology, including mobility in non-line-of-sight (NLOS) environments and terminals.
"QUALCOMM is committed to providing the best mobile broadband experience with 5G new port technology.We are driving the 5G mobile terminal, including smartphones, to become a reality in the world," said Christopher Amon, executive vice president and QCT president of Qualcomm. Covering a wide band from 6 GHz to millimeter waves and we are pleased to be working with Nokia to conduct trials based on the global 3GPP 5G standard, which is critical for the timely deployment of 5G networks.
"The consumer and communications service provider's interest and demand for 5G technology has become clear," said Marc Rouanne, president of Nokia's Mobile Networks business unit. "We and QUALCOMM are firmly convinced that the demand for large-scale mobile 5G is growing. , We are excited to work with QUALCOMM by using our 5G FIRST solution to develop and promote commercial technology and applications based on the 5G new port. "
The interoperability test will be launched in the second half of 2017 to keep track of the first 3GPP 5G new airport specification, which will be part of Release 15. The two companies will also join a number of network operators in 2018 in different regions 5G new air port OTA field trials, including China, Europe, Japan and the US Tracking the progress of the 3GPP specification is critical because it can facilitate compliance and verification of the global 5G standard, accelerating the introduction of standards-compliant terminals and infrastructure, and ensuring future 3GPP 5G new airplane version of the forward compatibility.
2, MediaTek to expand the supply Vivo will stabilize the market;
Vivo has been among the third largest smartphone brand in mainland China, mobile phone chip supply chain came a new product blueprint Vivo, now appears that in the first half of next year, MediaTek scored about three models.
From the Vivo product blueprint, the second half of this year will launch the use of MediaTek 'MT6750' (referring to product code) chip Y67 Next; next year were launched in the first half of the higher level of Xi Li (Helio) P23 chip Y54, and Using the first 12-nanometer chip P40 Y68.
Although Vivo flagship Xplay 7 and X13, X13 + and other flagship models still use Qualcomm chips, however, MediaTek won the number of models or more than this year, is conducive to market share in the next year, rebounded.
MediaTek this year by the smart phone chip product development behind the impact of the impact, resulting in almost the number of customers opened last year, the market share followed by a lot of decline, the annual smart phone chip shipments may be less than 400 million units.
3, Zhang Weili: Media Technology to follow the mainland market to grow together;
Mainland China, September 13, Beijing reported that 'the mainland market is broad, competitors must exist, an industry if there is no competition, on behalf of no living water. MediaTek is willing to see the competition, but also courage to participate in the competition in this process, We have to play their respective advantages, to provide customers with better chip products. 'September 7, in the Media Technology Beijing company's product showroom, MediaTek China mainland area chief representative Zhang Weili looked at the reporter to show a related technology In the face of the first half of this year's gross margin decline, Zhang Weili said: 'The declining figure is to inspire us to do another self-promotion, motivate us to provide customers with better chips.' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' '' ZTE, Lenovo, Huawei, oppo, vivo are all partners of technology, like oppo, vivo these are our early functional machine customers, I am pleased to see these two, three lines of mobile phone customers with the constant struggle , And together with the development of technology, in today's growth in the world can be a number of first-class enterprises. Qualcomm is worthy of respect for the competition Hand, but MediaTek as an Asian company to the mainland market-based, more familiar with and close to the mainland market, we talk with their customers speak the same language, know how to quickly provide products in line with the mainland market, which is the strengths of MediaTek I also believe that we can catch up with the US competitors quickly, although we are still working on this road. 'MediaTek entered the mainland in 2003 and has six subsidiaries on the mainland. At present, the most important business is mobile phone chip. MediaTek mobile phone chip business development with the mainland communications industry, as well as the development of the mobile phone market is closely related with the development of the mainland, MediaTek has also been growing in the mainland these years of development, I feel the deepest is: the most close to the market people In 2005, MediaTek to function machine into the mainland mobile phone market, when most of the market was occupied by European and American manufacturers, '2005 before the phone is expensive and not easy to use, low cost, MediaTek to Function machine into the mainland mobile phone market, so that the mainland people can afford to afford the price of the phone. 'With the function Machine market gradually saturated, coupled with the launch of the high cost of the chip, the mainland many manufacturers are using the MediaTek chip, more competitors also followed, so in 2009, MediaTek encountered A relatively large market challenges. MediaTek at different stages have encountered different challenges, but can constantly adjust their strategy to quickly launch in line with market demand for new products from the machine to the intelligent machine is a very important change. 2012 launched the launch of intelligent technology chip, which in line with the rapid popularization of 3G wireless communications technology in the mainland, so in 2014 we made a major breakthrough in the 3G market in the development process, we with the mainland operators China Mobile, China Unicom , China Telecom has also established a good relationship of cooperation, which also laid us with these operators to 4G after the development of the road more closely together. I am pleased that in 2G, 3G, 4G era, MediaTek can participate in the mainland Chip development, we are also happy to see in the future 5G market, MediaTek can also continue to grow with the mainland market. There are recent Internet of Things, a total of Cycling, we also have to contribute their chip products, promote industrial development environment. 'Zhang expressed peacekeeping force.
From 2001 to the mainland development, Zhang Weili said he witnessed the rapid development of the mainland economy, he said that when he first came to the mainland, from Zhongguancun Street has been to the north to see a lot of people riding a bicycle, today also see a lot of People riding a bicycle, but the original bicycle and the current bike is completely different, and now people riding a bicycle trip is to promote environmental protection, and now the bike has a high-tech elements, plus positioning chip, navigation chip, 'only from this At present, the mainland economy is facing transformation and upgrading, high-tech industry will play an important role in this process, Taiwan enterprises in the past few decades of development Focus on the technology industry, and the rapid experience of products, as well as the experience of dealing with Europe and the United States, management thinking is also more advanced, therefore, MediaTek as a high-tech enterprises, the future development in the mainland can be expected.Second, the mainland There is a vast market and excellent human resources, therefore, from the perspective of cross-strait industrial exchanges, just can learn from each other, I think this is whether it is The development of science and technology on both sides of the strait is also good for both the people on both sides of the strait.
4, the global semiconductor equipment market is expected to grow 37%, South Korea first, mainland China second;
According to SEMI's latest estimate, 2 global semiconductor equipment sales are expected to close to $ 55 billion in 2017, up 37% from 2016. This is the largest increase in global semiconductor equipment sales in 2017, But also since the financial tsunami since the highest growth rate.
SEMI further analysis, the world's largest semiconductor equipment market in 2017 will be won by South Korea, the value of its procurement of semiconductor equipment will reach 19.5 billion US dollars, a substantial growth of more than 130% in 2016. South Korea's semiconductor equipment purchases surge, mainly with Samsung (Samsung) Large investment in new memory capacity.
On the other hand, China's semiconductor equipment procurement is expected to surpass China's Taiwan in 2018 to 12.5 billion US dollars ranked second in the world, South Korea will maintain its first position in the world According to SEMI estimates, 2018, including mainland China, The United States, Japan, Europe, the Middle East and other places of the semiconductor equipment market, will appear more than two percentage growth, as other regions, the growth rate of less than 10%.
5, Heraeus Electronics launched the world's first non-splash solder paste and a new sintered silver;
Hewlett-Packard recently announced the launch of two new solder paste to meet the stringent requirements of semiconductor and power electronics applications, both of which are the leading materials solutions provider in the semiconductor and electronics packaging industry. Series of detailed technical information will be September 13-15 at the Taiwan International Semiconductor Exhibition (SEMICON Taiwan) on the disclosure, welcome to Taipei Southport Exhibition Hall, No. 2546 Heraeus booth. Heraeus Electronics is the world's first engaged in Splash solder paste commercial production of the enterprise, and this is about to debut Taiwan International Semiconductor exhibition of solder paste product line, then fully respond to Heraeus in providing customers with high-performance products, has been spare no effort.
WS5112 series solder paste is Hewlett-Packard's latest innovative water-soluble solder paste, with unparalleled performance, to meet the various processes and applications of demanding requirements. WS5112 is a water-soluble halogen-free solder paste, specifically Designed to reduce product defects and increase throughput. Due to its inherent excellent rheology, it can effectively prevent splashing and long service life, resulting in excellent fine pitch printing performance. The WS5112 flux platform supports a wide range of alloys And the highest No. 7. In addition, the solder paste not only has excellent wetting performance, and the residual residue is very low, it is easy to clean.
Hewlett-Packard's electronic mAgic® pressureless sintered silver can be described as "its name," which, like magic, can easily extend the life of the power electronics module and increase its thermal performance. ASP295-09P9 mAgic® Sintered Silver Lead-free, halogen-free formulation for efficient chip bonding solutions for power discrete devices, HP LED devices, RF power devices, etc. The sintered silver has a number of significant advantages, including: • sintering Low temperature (200 ºC) • Higher thermal conductivity compared to solder paste with higher heat dissipation and power density • No need to clean the flux after processing to reduce processing costs
Dr. Michael Joerger, head of global innovation at Heraeus Electronics, said that Taiwan International Semiconductor has brought together companies from different market segments where Heraeus is proud to showcase the roadmap and its latest results. 'Whether our customers are There are brand manufacturing, or from the field of equipment or materials, our goal is only one: to provide the perfect performance of products to meet the most demanding environment, application and process requirements.In order to achieve this goal, we invested heavily in building global R & D Network, and our launch of the water - soluble solder paste and mAgic ® sintered silver these two product lines is the best interpretation.
During the exhibition, Hershey Electronic Global Product Manager Chen Lishan will be the advanced system-class package solder solution for technical lectures. Venue: 3120 booth, TechXPOT 1F (material area); speech time: September 14 (Thursday) 1: 00-1: 30 pm.
6, the establishment of deep cognitive rules model artificial intelligence to break the future world blueprint;
AI development will be a combination of cross-cutting talent, covering biotechnology, medicine, engineering, science, psychology, management, etc., while AI can replace the human work, but also for our work, society, industrial structure brings many new AI applications will be blossoming, users will use and rely on high accuracy of product services, and enterprises will use artificial intelligence quickly from the digital era to upgrade to the intelligent era.
(AI), Baidu, Softbank, Facebook, and so on, the future of AI will penetrate into the life of the various aspects of Google, Microsoft (Microsoft), Baidu, Softbank (Softbank), Facebook And other international manufacturers, have to spend huge amounts of money to start AI research and layout.According to market research institutions Tractica's forecast, the global enterprise application of the AI market size, from 2015 to 2.025 billion US dollars, growth to 2024 years 11.1 billion US dollars , The growth rate of nearly 10 years close to 450. The recent artificial intelligence and depth of learning and other technology development heat warming, but also to promote the development of intelligent robot an important kinetic energy.
John McCarthy, a professor at the Massachusetts Institute of Technology, who has the 'father of artificial intelligence', defines AI as "science and engineering for intelligent machines, especially smart computer programs" 50 years ago. The combination of talents, medicine, engineering, science, psychology, management, etc., while AI can replace the human work, but also on our work, society, industrial structure brings many new challenges.
International manufacturers to preempt the layout of AI
Artificial intelligence is not a new term, as early as in 1956 in the Dartmouth meeting, 'Turing test' article formally defined AI, but so many years of its technological development experienced several ups and downs, the policy will MIC industry analyst Han Yang Ming (Figure 1) that the current common artificial intelligence technology is based on weak artificial intelligence (Weak AI), in order to help people a variety of behavior and decision-making, weak artificial intelligence does not emphasize the machine's self-consciousness, which can simulate people In contrast, the strong AI is strong AI, which emphasizes intelligent and conscious thinking and action, but this technology is not the focus of the current development of artificial intelligence, but it is not the same as the artificial intelligence The
Figure 1, MIC industry analyst Han Yang Ming pointed out that the current common artificial intelligence technology are weak Weij AI artificial intelligence, does not have the machine's self-consciousness.
AI over the past two years to come back, through the AlphaGo beat the chess king, to dominate the occasion, quickly become the subject of investment in science and technology, including Google, IBM, Baidu, etc. Han Yang Ming said, Google development of artificial intelligence technology for public use, Deepen the user's dependence on their services, and accumulate more information on sophisticated artificial intelligence, and then the development of paid services to provide business customers, but also continue to merge with a special ability to start a new company and team to strengthen their overall strength.
One of the industry's most famous artificial intelligence IBM Watson, in recent years, actively to the medical field development, the introduction of personal medical guidance program, electronic health record analysis, blood glucose management program, artificial intelligence medical plan, etc. also developed clothing design, And other professional areas of services to help enterprises for intelligent applications, and self-accumulated data interpretation capacity.In addition, Han Yang Ming, Baidu focus on independent research and development, and in the data collection has a sufficient geographical advantage to enhance the training of information after the service Precision, while large-scale investment in artificial intelligence efforts to transform, the development of experimental intelligent hardware products, its voice service has a world-class level, the recognition rate of 97% Mandarin, and the development of intelligent enterprise cloud services, (O2O) intelligent service battlefield.
But the wide range of artificial intelligence applications, not only the manufacturers can play, Han Yangming that a variety of platforms and open resources more and more, so that small and medium enterprises and even individuals can quickly cut into the intelligent service; the future is not just a use Artificial intelligence technology research and development, but the use of composite technology to support each other to create more and more effective products / services; AI applications will be flourishing, users will use and rely on high accuracy of product services, and enterprises will use Artificial intelligence quickly upgraded from digital to intelligent times.
AI from the cloud to the terminal
AI is a hot topic in recent years, but it is not so far away, ARM mobile communications and digital family senior market manager Lin Xiuping (Figure 2) pointed out that according to the groundwork letter (Deloitte Global) research, 2017 more than 300 million Mobile phone with AI capacity to the current annual market size of 1.5 billion mobile phones, there are 1/5 of the phone with AI function, these consumers often use the features include: indoor navigation (Indoor Navigation), the expansion of reality (Augmented Reality), translation, image classification, speech recognition, etc. The application domain will gradually migrate from the cloud to the terminal, especially for personal or security-related applications.
Figure 2 ARM mobile communications and digital family senior market manager Lin Xiuping said that in 2017, 1/5 of the phone with AI function.
Action AI application can be divided into two categories: First, the man-machine interface, one is the system / security related to the future application of AI more and more widely, the microphone on the phone and the camera plays a whim function, carrying the number will continue to improve, the future A mobile phone equipped with a camera lens up to 8. In response to AI function mobile phone computing power will continue to improve, CPU is responsible for dealing with logic operations, GPU is responsible for image processing parallel operation, processing capacity will continue to improve; AI's development needs have gone up, including passive security protection and active security features.
ARM as a CPU IP provider, for the application of AI, the future will continue to optimize its product architecture and computing power, Lin Xiuping pointed out that the company continued the past big.LITTLE architecture, introduced DynamIQ architecture, the next 3 to 5 years can be improved CPU performance than the existing 50 times the AI, including Compute Library average 4.6 to 15 times the performance improvement, with the core 3 to 5 times the performance improvement, integrated down to improve the performance prediction.
Deep learning is the core of AI technology
AI has recently been highly regarded by all walks of life, because the technology is about to bring about significant changes in people's economic activities, but in the past these innovations for people to save the labor, this is the brain recently discussed AI may replace a lot of human Of the work, which caused the unemployment problem, in fact this is a comprehensive promotion to help mankind from the original boring repetitive work freed, DeepBelief.ai chief artificial intelligence scientist Yin Xiangzhi (Figure 3) that AlphaGo For example, the computer is still not thinking or self-conscious, and there is no strategic ability, he is only in the Go this activity, through the depth of learning to find the best winning formula, while a large number of data to learn and get result.
Figure 3 DeepBelief.ai chief artificial intelligence scientist Yin Xiangzhi that the computer so far still no thinking or self-consciousness, there is no strategic ability.
Simply put, the difference between man and animal is that so far we are the only one with deep perception of the creatures, Yin Xiangzhi stressed that we can understand, induction deep rules, but can not be specific to this ability, so now the depth Learning is to 'imitate' this kind of neuron operation, the human intuition into a logical rule, which has three of the most important schools, one Convolutional Neural Networks (CNN), the main application In the machine vision; the second is the recurrent neural network (Recurrent Neural Network, RNN), mainly used in the language model to establish; third is Reinforcement Learning (Reinforcement Learning), the most typical case is AlphaGo, is a control mechanism.
Image analysis is probably one of the best technologies developed in the past few years, through the characteristics of the clustering, from the pixel level to find the characteristics of the convolution of the neural network is the matrix of the pixel operation, the characteristics of stratification And then layered, to identify the characteristics of the characteristics, and finally can achieve our deep perception of the effect, such as through a photo to recognize a cat, can be applied to identify classification, target detection, semantic segmentation, action detection.
Language model from the development of voice recognition, 2016 Microsoft voice recognition accuracy rate officially more than human professional stenographer, Google translation is also used to recursive neural network model, basically in the language part of the language, English-based European and American languages are not too Big problem, but the Chinese is still a bottleneck, Yin Xiangzhi said, there are more than 20,000 words in the light, flexible language structure, including Chinese and English mixed, dialect, buzzwords and so on for the language data to enhance the complexity, and through Natural language to organize the knowledge map, knowledge map and then through the natural language reasoning, AI is the next wave of development focus.
Improved learning is to interact with the real world to find the best combination of strategies to maximize the value of AlphaGo training computer how to find the best strategy, Yin Xiangzhi said, DeepMind is to strengthen the learning to let artificial intelligence to visual interaction to play video games , And to strengthen the value of the establishment of the computer to understand the winning game is the value of the direction of the computer to continue to establish a strategy to win the development of the game, the forecast model is to predict changes in environmental status and predict the impact of changes in value, and then continue to respond, Win the strategy / practice.
AI has become one of the core elements of national competitiveness
AI development and technology to master, but also the national competitiveness of the show, so in addition to industrial investment, the world's science and technology power also dumping the power of the development of the State Institute of high-speed network and computing center deputy researcher Pan Yilun (Figure 4) The government released the future 'National Artificial Intelligence Research and Development Strategy Plan' and 'ready to meet the future of artificial intelligence' strategic report, the AI development at the national strategic level, while Japan perennial efforts to develop robots and AI, Japan's new energy industry technology development agencies NEDO) announced that Japan's next generation of artificial intelligence social use vision will be manufacturing, action life, medical / health / care, wholesale and retail and circulation of four exports as 2025 ~ 2030 to achieve the focus.
Figure 4 State University High-speed Internet and Computing Center, deputy researcher Pan Yilun said that Taiwan must be upgraded from hardware to upgrade to innovation and intelligent software research and development to a place in the AI industry.
In recent years, the development of the latest technology in mainland China, but also in 2016 China Development and Reform Commission and the Ministry of Science and Technology announced the 'Internet + artificial intelligence three years of implementation of the program', the artificial intelligence included in the thirteen five plan, AI will be promoted to one of the strategic industrial projects The South Korean government also announced that it would invest about US $ 27.68 billion by 2020 to promote the research and development of the artificial intelligence industry and set up a research center by the people as a national AI industry research and development hub, and Taiwan, including Kai-fu Lee, Jane Everest, Du Yijin And so have made recommendations on the development of Taiwan AI, but one of the consensus is that Taiwan in the past too heavy hardware and software, which is the development of AI crisis, Taiwan must be upgraded from hardware to upgrade to innovation and intelligent software research and development, And strengthen the integration of soft and hard talent and the market, at the same time with the top academic research and domestic manufacturers of hardware manufacturing capacity to obtain a place in the AI industry.
Model builds like traditional skills
To import AI to first set out the technical architecture, the industry has a lot of related tools, Taiwan University of Electrical Assistant Professor Li Hongyi (Figure 5) said that in general AI model is divided into two parts is learning or training, the other is the test. There are three steps in the architecture of learning or training, the first is to determine the structure of the network, especially in the depth of learning architecture, we must first determine how many layers of the network, each layer of the number of neurons, the second step It is necessary to define the goal, that is, the focus of the structure, for example, is to identify the cat, we must first provide a lot of data to tell the characteristics of computer cats in order to its accurate identification; the third part of the algorithm to choose, and then with the fastest Way to find the best solution.
Figure 5 Taiwan University Department of Electrical Engineering Assistant Professor Li Hongyi explained that the establishment of the depth of learning model If you choose the tools, pour the appropriate information, it will be very accurate to complete the goal.
In the above process and results, some very simple and very difficult skills, similar to the traditional handicrafts, Li Hongyi explained that the establishment of the depth of learning model is like making a bad pull, hand-drawn embryo master can use simple raw materials, Lightweight approach, smooth completion of a perfect work, but we do not necessarily do so; the depth of learning model if you choose the tools, pour the appropriate information, will be very accurate to complete your goal, but if not enough to understand Depth learning technology architecture and model building methods, will encounter great obstacles, which is the difference.
AI application to create new business opportunities
AI from the earliest proposed in the 1940s, after several times the development is not smooth, the policy will be senior industry analyst Li Zhenhua (Figure 6) pointed out that in recent years, large-scale growth of various data, artificial intelligence technology gradually mature, The rise of large data, the expansion of cloud computing and storage technology and the rise of Internet of things and other factors, to take advantage of a breakthrough in the future growth in the industrial applications, including manufacturing, circulation, the financial industry are popular applications.
Figure 6 will be senior industry analyst Li Zhenhua MIC pointed out that Taiwan's development of AI, to lock the priority demand field.
A large number of AI applications, financial robots is a very popular one, the past two years, robot managers manage the rapid growth of assets, is expected to 2015 20 billion US dollars, grow to 2020, 450 billion US dollars, showing the robot financial technology Mature, the market acceptance is getting higher and higher.In addition, the US financial giant JP Morgan Chase & JP in 2016 launched contract analysis intelligence software COIN, instead of lawyers and credit officers to review the contract documents, 10,000 hours of work, shortened to a few seconds to complete.
In the process of developing AI, Li Zhenhua proposed to lock the priority demand field, such as intelligent factories, long-term care, intelligent transportation / transportation, etc .; In addition, to focus on specific niche areas, such as Chinese language processing, computer vision, video Identification, etc .; the last is the combination of Taiwan's industrial advantages, such as IC design, foundry industry, security control industry or system integration, combined with AI technology, the development of specific areas of application services, key applications or complete solutions.
AI Patent Layout and Analysis
AI has recently been widely discussed in the industry, the patent layout has become the focus of the layout of the country and the major manufacturers, the policy will be senior industry analyst Chen Cixian (Figure 7) analysis, for 22,976 artificial intelligence US patent international rights of the country For the analysis, it found that the United States holds 61.9% of the patent, accounting for 11.3% in Japan, the total up to 73.2%, compared with Taiwan only 1.0% .In addition to the field, the main focus on computing technology up to 76.8% Followed by control, measurement, digital communications, information technology management methods, telecommunications, audio-visual technology.
Figure 7, MIC senior industry analyst Chen Zhixian refers to the analysis, Taiwan is currently in the field of artificial intelligence patent layout is relatively backward, and focus on natural language processing, adaptive control system.
Taiwan is currently in the field of artificial intelligence patent layout is relatively backward, and focused on the natural language processing, adaptive control system, Chen Zhixian description, intelligent customer service is currently a small number of patents and wide range of easy to find niche market, is the need to develop new Application of the market.It is recommended that Taiwan's technology industry should first lock the specific application areas, such as the management of office automation, business class customer management, financial class credit, etc., to further build the domain knowledge base or expert system and develop the corresponding voice Identification, learning methods, machine learning, natural language processing, and do a good job with patent applications and layout planning.
In addition, the number of patents related to the learning method is not much in the related technologies of the intelligent image detection patent. Chen Zhixian suggested that Taiwan's related industry could devote the research and development of the research and development resources, Learning, enhanced learning and other related technologies in-depth study, while doing a good job in foreign countries, especially the US patent application and layout, and actively protect Taiwan's intelligent image detection of independent innovation technology.