Set micro-grid launch micro-channel IC WeChat public number: 'Daily IC', real-time release of major news, every day IC, every day set of micro-network, micro-into a!
1. NVIDIA Huang Renxun: This is the beginning of the AI revolution, and we are in it.
The future of artificial intelligence in the ubiquitous, from autonomous driving, robotics, smart city to medical diagnosis.In addition to competing on the platform of Google, Microsoft, Apple and Tesla and other technology giants behind prop up this new millennium, is this Born in Tainan, nine-year-old Chinese entrepreneur to the United States - NVIDIA founder and CEO Huang Renxun.
On the morning of October 26, less than ten o'clock, the second floor of the Le Meridien Hotel in Chicago was filled with crowds of people, and the restlessness and excitement in the air were obviously more refreshing than the coffee. This is the GTC 2017 meeting place in Taiwan. All are concerned with artificial intelligence and in-depth study Developers, researchers and industry leaders are here, trying to capture the future of the world from the words of Jen-Hsun Huang, founder and CEO of NVIDIA.
Light up, always wearing a dark leather coat appeared in public Huang Jen-hsun, a jogging dress on stage. "This is the beginning of a new era," he said: "This is the beginning of the AI revolution, And we are in the midst of it, riding on the waves. "
Indeed, in the past year, NVIDIA can be said to be standing on the highest wave, according to the latest earnings report, NVIDIA's revenue and earnings per share hit a record high, exceeding expectations, while the market is optimistic about this year, NVIDIA's stock has risen 92%, its market cap more than 100 billion US dollars, double the previous year and become the most watched industry superstar, all of which are closely related to NVIDIA's field of artificial intelligence in recent years.
At this year's GTC conference, Huang Renxun has talked about everything from autonomous driving, Internet of Things devices, virtual reality to medical diagnosis, and even announced that it will co-operate with the Ministry of Science and Technology to jointly train 3,000 developers over the next four years to incubate Industrial Ecosystem, which also means that NVIDIA will not be absent in various future scenarios.
Nvidia can never be described as smooth today, but in particular, Huang Renxun can see farther than anyone else on every strategic turn and secretly prepares gunpowder to meet the challenge until the next wave of trends.
NVIDIA's first crisis burst just after the company was founded in 1995, NVIDIA spent two years of research and development, finally introduced the first display chip NV1. However, this chip along with the subsequent launch of NV2, but because with Microsoft Windows 95 development of the Direct3D standard is not compatible, and therefore can not get into the market.Huang Jen-hsun, the wrong bet, had decided to layoffs, the number of companies from more than 100 people cut to more than 30 people.
In order to return to the main battlefield, Huang Renxun introduced the technology of new blood, digging corner specialized high-performance computing Dr. David Kirk as chief scientist and in 1999 the GPU GPU family of products introduced GeForce256, Solidified NVIDIA's position in the computer graphics industry and, under Cork's leadership, NVIDIA introduced the CUDA Compute Unified Device Architecture, which uses GPU processing power to increase computing performance. , Is a major turning point that affects the later development of NVIDIA.
GPU development, not only to bring the wave of artificial intelligence revolution, on the other hand, NVIDIA also opened the switch from the display adapter hardware vendors to the opportunity of artificial intelligence platform company in 2012, University of Toronto graduate Alex Sikowski ( Alex Krizhevsky) Within two days, using two NVIDIA GeForce GTX 580 GPU training neural network computing logic AlexNet to identify images and win the ImageNet contest with this essay, the message immediately caught the attention of Global Artificial Intelligence researchers Since then, many researchers of artificial intelligence began to use NVIDIA GPU deep learning.
Now, in web services, transportation, healthcare, finance and manufacturing, researchers in every major industry have used NVIDIA to develop their own artificial intelligence, and even the Google artificial intelligence AlphaGo and Tesla electric cars that defeated World Chess Kings Is GPU-based computing, and so far more than 2,000 AI startups have been built on top of NVIDIA, making NVIDIA the founders of the truly artificial intelligence era. "I think in a growing company As CEO, we must constantly remind ourselves: every few years, I will reform myself, and do not be afraid to make mistakes. "From the current results we can see that Huang Renxun's self-reminder, not just to talk about it.
Aiming at the trend, bold attempt, continuous evolution. In the time of the quench, NVIDIA has long been out of the prospect of as a creator, the past few years, Huang Renxun personal mood changed? "No different, I think Like a teenager! "The same black leather, the same confident look, the future, his self-combat will continue. The digital age
2. Microsoft Asia Research Institute: AI to data as the core of application innovation welcome innovation;
In recent years, the issue of artificial intelligence has repeatedly become the focus, and Michael Jordan, a professor at the University of California at Berkeley and a fellow of the U.S. Academy of Physicians, said that today is not in a magical big bang era of artificial intelligence, it takes another hundred years to put this high-rise It is not a simple algorithm to build, but to create a market that combines the needs of users at both ends of the enterprise, allowing interaction and operation in this area to be easy and then adjusting the learning step by step so that artificial intelligence can change Decent intelligence, the evolution of the IT Internet over these years, shows that demand-side deepening of applications can in turn lead to technological advances that are no exception in the area of artificial intelligence, but all the requirements exist at the forefront of every industry and how these The combination of needs and cutting-edge artificial intelligence technologies is exactly what Microsoft Asia Research put into thinking in the past year or two.
At the present stage, the core of AI development is data. However, in the various industries, the accumulation of a large amount of data on the front line, if the effective use, it will bring about the change of the industry. In the opinion of Microsoft, academic research in AI field But also must be closely integrated with various industries, the only way to hit the spark of cross-border innovation, and in Wall Street, machine set the trading point of stock futures, automated trading system is a new normal, AI has long been the core of financial enterprises to enhance Competitive killer technology; services such as street corner, internet banking or community bank service and personal consumer loan repayment are also facilitated by the infiltration of AI. The common denominator is the use of large amounts of data to analyze revenue nodes and financial risks .
Zhang Yizhao said the vast wealth of data and information is the biggest gift of this era, and they are also subverting the most common financial scenarios.For example, any data of individual lenders may reveal key information, even if the release of friends Circle, sports bracelet positioning, can become a source of information in the face of emerging financial management, insurance new products, in the end which is most suitable for their own, may be a professional financial advisers, insurance consultants may not be able to immediately give the most appropriate The answer, one end is the explosion of data and information, and the other is the analysis, summary and transmission of valid messages, the message divide between users and service providers, is the AI's potential entry point in the financial industry.
Zhang Yi Zhao also said that the machine is better at data memory and finishing, to make it more intelligent and have more analytical capabilities, it is the current stage of the AI goal, in addition to different clients recommend suitable financial, insurance and other financial Products, but also provide thousands of financial services, and in the financial industry's back-end, AI power of the same huge space, similar to 'interest rates rose 0.5% over the past 50 years the entire market how the impact?' For human beings, it may be clear that only senior analysts can explain it clearly, but it can quickly analyze analogy models in similar historical situations by comparing timeline data, and even analyze and predict. The difference between the two is that analysts To come up with such a model, it may take decades of experience, and AI can quickly get through the boundaries of the field with the help of industry experts.
He further pointed out that a few years ago the tsunami in the Indian Ocean caused floods in many cities in Thailand. What is unknown is that the tsunami hit not only the homes of ordinary people but also the global PC shipments This is because many PC parts are manufactured in Thailand. Generally speaking, the relationship between the tsunami flood and PC shipments will not be easily noticeable, but it actually exists. For the covert internal relations , Industry analysts can use the AI and big data to insight, and make the appropriate judgments.
Similarly, the devastating floods that occurred in Texas this year led to the shutdown of many refineries. This certainly affected oil prices. Due to the complexity of the influencing factors in the upper, middle and lower reaches of the entire eco-system, Analysis, the refinery is in the middle of the entire ecological chain, if the stop production will not only lead to lower prices of refined oil, the upstream crude oil is also likely to sell poor, resulting in price decline, while in the real world, there are many similarities, How to include more variables, to build a more reasonable model in a broader context, are all issues that Financial AI is exploring.
In June of this year, Microsoft Asia Research Institute and Huaxia Fund announced that they would conduct strategic cooperation research on the application of artificial intelligence in the field of financial services. The research directions include forecasting market trends through pattern recognition, mining important factors influencing the market based on deep learning, Learning Methodology Research on industry rotation, building financial graphs based on big data, usage data based on social networks and application software, identifying and in-depth understanding of customers, etc. Insiders hope to take this opportunity to study the AI + financial frontier and promote the financial sector Intelligent transformation, and let massive data and information really value, it is the mission of the financial AI where analysts used to be often barely as strawberries, now rich in raw materials, clever women can take advantage of more efficient AI tools to make the data better, the analyst's responsibilities will be adjusted accordingly to translate and interpret the results of the AI analysis based on years of industry experience, and then to verify and optimize before building Model; AI + financial is not a human replacement, but re-division of labor, AI + HI Together, the entire industry will greatly enhance production efficiency.
Zhang Yi Zhao pointed out that in the Microsoft Building, Beijing, there is a new vending machine, which comes from Microsoft accelerator Gan company, the appearance of vending machines and roadside vending machines look very much the same, but it's Glass cover hidden mystery, closer look will find that the glass cover is also a screen, like the machine's eyes, you can identify the shopper's gender, age, and even when picking the expression of the goods, if in front of it for a while, You may find you take the opportunity to recommend to customers some may be interested in trinkets and so on, and when customers brush credit card, WeChat, Alipay checkout, consumer habits, interactive records are stored, which for retailers and Words, is the freshest, valuable, real first-hand data.Zhang Yizhao also pointed out that this vending machine can be considered a prototype of AI + retail, a lot of AI technologies such as face recognition, computer vision, natural language processing, etc. can Apply to each supermarket checkout, businesses do not need to identify the characteristics of a specific customer, you can get the entire consumer groups atlas, like NIKE, ZARA, STARBUC KS and other famous brands, in recent years are also increasingly concerned about the evolution of the characteristics of its customers and potential customers, the past to go through the staff of copycat, investigators to do the work together in the future with the AI, the related work will be more simple, Accurately and efficiently, to better align the brand tonality.Moreover, issues like how consumers like their products, how much inventory they should stock, etc., will no longer be a problem with the help of retail AI, such as: The rise of unmanned stores and the forecasting of masks, air purifiers sales and stocks through weather forecasts can all be analyzed and predicted more and more accurately with the help of artificial intelligence.
Zhang Yizhao said that in addition to the financial and retail industries, the logistics industry is one of AI's rapidly infiltrating industries. Whether it is sharing bicycles around you, or the logistics system behind couriers / takeovers, or overseas purchases, the global trade involves The global logistics system has undergone tremendous changes under the influence of artificial intelligence. Even with the background of some logistics enterprises, as the volume of data increases sharply, the industry insight they have gained far beyond what people think.
Zhang Yi Zhao also said that another full of futuristic scenes occur in the manufacturing industry, in fact, Industry 4.0, Made in China 2025 can be seen as AI in the manufacturing industry's vision for the future, more and more intelligent robotic arm , Robots and sensors, have been deeply applied in many scenarios without too much interaction with humans. Several companies that are incubating Microsoft accelerators have also made helpful attempts in this regard, such as the use of unmanned aircraft technology to explore power generation wind turbines , High-voltage cables and other special equipment materials, cracks, and the corresponding maintenance work, etc. In contrast, in the logistics, manufacturing and other end-users do not have too much interaction scenarios, the application of artificial intelligence rather fast However, in the early days, the fields of medical treatment and health that everyone favored, however, have been slow to apply artificial intelligence because of the restrictions imposed by various management mechanisms and human interactions.
Zhang Yizhao pointed out that in addition to seeing the impact and changes brought by AI to various industries, as the forerunner of artificial intelligence research, Microsoft Research Asia naturally hopes to further promote the deepening of the AI + industry in the era of artificial intelligence, Accelerate the arrival of the era of artificial intelligence, so, with the idea of establishing innovation exchange, hope innovation exchange As a convergence of Microsoft Research Asia top technology experts and experts in various industries innovation wisdom, experience, technology platform that allows Microsoft Research Asia and Large-scale enterprises, investment institutions and government departments in the Chinese market will establish more extensive and in-depth cooperation ties and build a cross-industry communication platform for all to discuss technological innovations in the AI era.
In addition, Zhang Yi Zhaoxi also pointed out that although Microsoft Research Asia with leading AI technology and strong scientific research strength, but researchers for real industry scenarios do not understand, there is no real industry data, based on a small amount of data It is time to build a model and the future of AI must be based on big data, so the combination of AI and the industry is an inevitable trend, and now it is precisely the best time to start.He further pointed out that to make AI more Grounding gas is the correct idea of rapid development, in order to speed up the conversion of AI into productivity, the real landing speed.
Zhang Yi Zhao said at the same time, for Microsoft Research Asia, will continue to basic research in the field of computers, while the creation of innovation is to enable researchers and researchers to better consider from the practical point of view Scientific research issues, the future will be through closed-door technology conferences, seminars and other forms of bridging the wisdom and industry resources, specific problems encountered in the digital transformation of enterprises, tailor-made a dedicated part of the technology exchange, and a business together Explore the bright future of AI + industry
3. German machine vision plant Isra Vision jumped 10%, a record high;
Isra Vision AG, the German provider of 3D machine vision technology, today announced fiscal 2016/2017 fiscal year 2016 revenues: revenues rose 11% to € 143 billion, gross profit margin rose to 57% from 56% a year ago, EBITDA (before tax, EBITDA rose 14% to 42.8 million euros, EBITDA rate rose 29% from 29% to 28 million euros, while earnings per share rose 17% to 4.68 euros.
Isra will release full fiscal 2017/188 results in February 2018. Isra's current backlog is more than 90 million euros, up from 85 million euros a year earlier. The company initially estimates a profit margin for 2017/2018 of at least Steady at last year's level, revenue is expected to grow 10-14%.
As of September 30, 2017, the cash / cash equivalents of Isra Vision increased by 76% to € 29.7 million and the inventory was reduced by 3% to € 32.7 million.
Thomson Reuters quotation system shows that IsraVision shares in Germany rose 10.66% on the 15th to close at 199.75 euros, a record high; the cumulative gain over the past year reached 103.52%.
QY Market Insights pointed out in September that 3D machine vision vendors include Keyence Corporation (6861.JP), Cognex Corp. (CGNX.US) and Isra Vision.
Harvest XQ global winners quotes system, Keyence hit a record high of record closing on November 24, so far this year (as of December 15 closing) cumulative increase of 57.9%.
Cognex, the global leader in machine vision for process automation, rose 0.13% to $ 60.97 on the 15th, up 91.67% so far this year.
Cognex realized 76% revenue growth (up 50% quarter-over-quarter) in the third quarter of 2017 (up to October 1, 2017) to a record 257.973 mil USD; its gross profit margin dropped to 76 from 78% a year ago %; Profit margin rose to 42% from 37% a year ago; non-GAAP diluted earnings per share jumped 91% to $ 1.11 from $ 0.58 a year earlier.
Cognex founder Robert J. Shillman pointed out that Cognex's Q3 revenues, net income, earnings per share and profit margins all set record highs.
4 technology giant to seize the AI market, how smart chips bring change to life;
Technology giant giants seem to have fully embraced the AI revolution, and Apple, Qualcomm and Huawei have created a mobile chip that is designed to provide machine learning a more Good platform, and different companies have adopted a slightly different approach to the design of this chip.ChinaIn this year's IFA released the Kirin 970, they call the first chipset with a dedicated NPU (NPU) Apple then unveiled the A11 bionic smart chip, which powers the iPhone 8, 8 Plus and x. The A11 bionic chip features a Nerve Engine processor designed specifically for machine learning.
Last week, Qualcomm released the Snapdragon 845, which delivers artificial intelligence tasks to the core processor of the most suitable processor, and the design approaches for the three companies are not that different - ultimately boiled down to offering developers The access rights provided, and the amount of power consumed by each setting.
Before we discuss this issue, let's first figure out how an artificial intelligence chip differs from an existing cpu. In the industry, you often hear the term "heterogeneous computing" about artificial intelligence. Is a system that uses multiple processors and each has its own specialized features for higher performance and energy savings.The term is not new and is used by many existing chipsets - For example, these three new products have adopted this concept to varying degrees.
For the past three years, smartphones have used the ARM big.LITTLE architecture for cpu, which combines the relatively slow energy-saving core with the faster, less-energy-intensive core, and our main goal is to make this chip perfect May consume less power for better battery life.The first handsets to adopt this architecture include the Samsung Galaxy S4, which only incorporates the company's own Exynos5 chip, and Huawei's Mate8 and honor 6.
This year's 'artificial intelligence chip' takes the concept a step further by executing a machine learning task by adding a new dedicated component or using other low-power cores for machine learning tasks, for example the Snapdragon 845 can take advantage of it Digital signal processor (DSP) to handle long-running tasks that require a lot of double counting, such as finding a hot word for a user through analysis over a long conversation Qualcomm's director of product management, Gary Bulotman, told Engadget, On the other hand, needs like image recognition can be better managed through the GPU, with Brotman solely responsible for developing artificial intelligence and machine learning technologies for the Snapdragon Smart Platform.
In the meantime, Apple's A11 bionics application adds a neural engine to its GPU to speed facial recognition, verbal feedback and the use of third-party applications, which means that when you start these processes on iPhoneX , A11 turns on the Nerve Engine to perform calculations to verify the identity of the user, or pour your facial expression into the 'talkative' app.
In the Kirin 970 chip, the NPU handles tasks such as scanning and translating images in Microsoft Translate, the only third-party application optimized for the chip to date, and Huawei said its 'HiAI' isomerism Computing architecture maximizes the performance of most components of its chipset, so it may allocate AI tasks to more than just NPUs.
Despite these differences, this new architecture means that in the past machine learning computing was handled only in the cloud and can now run more efficiently on the device itself. By using non-CPU parts to run AI tasks, users' phones You can do more things at the same time, so you do not have to delay until you have the app translate or search for a picture of a pet dog, for example.
In addition, running these programs on your phone eliminates the need to send users 'usage data to the cloud, which gives users greater privacy because it reduces hackers' chances of getting data.
Another big advantage of these artificial intelligence chips is that they save energy, because some work is repetitive, and our cell phone battery consumption needs to be more rationally allocated for these repetitive processes.GPUs tend to absorb more energy, so if instead Is a more energy-efficient DSP, and it can achieve similar effects as GPUs, it is best to choose the latter.
What needs to be made clear is that the chip itself does not decide which core system to use as a driver in deciding to perform certain tasks. "Today, developers and OEMs want to run AI chips," said Brotman. Use the support database (or, more specifically, its Lite mobile version) like Google's TensorFlow to choose the core from which to run their models Qualcomm, Huawei and Apple all use the popular TensorFlow Lite and Facebook's Caffe2 Options as their design support program Qualcomm also supports the new Open Neural Network Switching (ONNX) system, while Apple has added compatibility for more machine learning modes through its core ML framework.
So far, none of these chips have had any real impact in the real world, and chipmakers are touting their own test results and benchmarks, but these test results are not until AI programs become an important part of our daily lives Are meaningless because we are in the early stages of developing machines for machine learning and there are very few developers who use new hardware.What the smart chip brings to life is how the tech giants grab the AI market
But now it is clear that competition has begun and competitors are focusing on how to make machine learning-related tasks run faster and more efficiently on user devices. We only have to wait for a while, Chip to artificial intelligence chip changes bring us life's help.
Photo: Huawei (Kirin AI processor), Apple (A11 processor core). (From: Engadget compiler: NetEase see the external compiler robot review: Fu Zeng)
5.Google Neural Network assists NASA in identifying new planets highlighting the limits of AI technology
The use of Artificial Intelligence (AI) has now been extended to space exploration, with NASA announcing the discovery of two extrasolar planets with the help of the neural network developed by the Google AI team, one of which is named 'Depler- 90i 'around the 2,545 light-year-old' Kepler-90 'stars NASA astronaut Andrew Vanderburg described the Kepler-90 as a mini-version of the solar system in which planets are arranged from small to large. After the planet, the Kepler-90 has eight planets equal to the solar system. In fact, two planets existed in NASA's Kepler satellite four years ago. Although NASA discovered them, they did not go deeper Kepler's mission is to look for planet-like planets outside the Milky Way and detect over 35,000 possible planets over the years, as it is too large to be manually screened. NASA has the highest possibility of being an automated system Of the planet, but sometimes weaker signals are missed.Now Vanderburg and Google engineer Christopher Shallue decided to use machine learning and Google's neural network to filter out old data, Fifteen thousand Kepler signals trained the network and then analyzed nearly 1,000 weak signals from systems known to have planets, eventually digging out two planets that were buried for many years. Vanderburg said the neural network is still not perfect and sometimes There are mistakes, but can find more planets that have been missed before. NASA's new findings underscore the potential of machine learning beyond computer science, and NASA said it will use AI for more exploration missions in the future. "DIGITIMES