On the eve of 5G commercial, AI function will become the best design of mobile phones in the future. Qualcomm plans to become the first choice of various application platforms, and has established cooperation with many partners. It is foreseeable that the application of AI will explode in the future. AI has injected a fresh blood into the "innovative and weak" mobile phone market. In 2017, the focus of the technology industry is not artificial intelligence (AI). Regardless of technology giants, Internet giants, investors, and terminal manufacturers, almost all companies have devoted themselves to the field of artificial intelligence. It seems that the era of artificial intelligence is coming. The concept of artificial intelligence has been proposed for more than 60 years. During the period, it has experienced two limitations, suffering from the limitations of hardware and software technology, and has not made much progress for decades. Nowadays, with the rapid development of parallel computing, big data, deep learning algorithms and other fields, artificial intelligence technology has advanced by leaps and bounds. As the world's largest consumer electronics consumer, mobile phones have a large user base, but almost all research data shows that in the next few years, the global mobile phone market will maintain a single-digit growth, on the eve of 5G commercial, if the mobile phone does not “Subversive†innovations can hardly motivate users to change their needs. The combination of AI and mobile phones has given new hope to innovative mobile phones, becoming the biggest selling point for mobile phone differentiation. In 2017, the latest mobile phone chips released by Qualcomm, Apple, and Huawei, the introduction of AI functions are the most invented places. The A11 chip of the new iPhone is built with the artificial intelligence "neural engine", which has been applied in face recognition; Huawei's latest Kirin 970 processor has a specialized neural processing unit (NPU) and is used in the release. Huawei's Mate 10 series and Glory V10; Qualcomm's Snapdragon 845, which was released just a week ago, is a distributed architecture-based neural network processing engine (SNPE) that runs on CPU, GPU, and DSP of the Snapdragon heterogeneous platform. Wait for every unit. It is worth noting that this time in terms of the implementation of AI functions, mobile chip vendors have taken a different approach. Both Apple and Huawei have chosen the most popular "Special Application Integrated Circuit" (ASIC) method to implement AI functions and have independent AI modules. Qualcomm used heterogeneous computing to achieve it. At the just-concluded second Snapdragon Technology Summit, Alex Katouzian, senior vice president and general manager of mobile business at Qualcomm Technologies, Inc., introduced in an exclusive interview, "Qualcomm The product architecture itself has the NPU's ability to support machine learning on different engines, including GPUs and DSPs, and has enough intelligence to distribute different use cases to different cores instead of adding additional NPUs. †Christiano Amon, Executive Vice President and President of QCT, Qualcomm Technologies, Inc. (left) Alex Katouzian, Senior Vice President and General Manager, Mobile Business, Qualcomm Technologies, Inc. (right) Since the release of the Snapdragon 845, many people have slightly succinct about AI implementation, and there is no independent AI module in Qualcomm. In fact, Qualcomm's exploration in the AI ​​field started very early. In 2007, Qualcomm began to launch artificial intelligence projects, and began to explore machine learning pulse neural methods for computer vision and motion control applications, and then extended its research from bionic methods to artificial neural networks. Cristiano Amon, executive vice president and president of QCT at Qualcomm Technologies, Inc., stressed that the Snapdragon 845 is already Qualcomm's third-generation artificial intelligence platform, and its overall performance is three times higher than the previous generation. Amon said that Qualcomm has been committed to ensuring that the Opteron platform has machine learning features and supports the operation of many mobile use cases, many of which may have been running in the cloud. Because mobile terminals have some different qualities in terms of power consumption and heat dissipation, we have built a very flexible architecture that allows machine learning use cases to run on multiple engines, such as GPUs and DSPs. Qualcomm believes that for the characteristics of mobile terminals, it can provide more flexible deep learning and coordinate the capabilities of each core, so it can achieve the highest efficiency and best performance in today's mobile terminal processing. The deep learning function is also expanding to the mid-end chip. Alex Katouzian said, "We found that most of the engines that do this kind of operation (called NPUs by other companies) work on vector processing, and these can be perfectly processed in our DSP. Qualcomm firmly believes that Independent components are hard to compete with integrated solutions. The integrated solution allows us to scale, and allows us to not only bring machine learning to the top-level platform, but also to a broader level, such as high-end éªDragon 600 series." We can understand that Qualcomm uses SNPE to coordinate the functions of various hardwares, so that CPU, GPU, and DSP all assume the computing functions of AI, and different tasks are run by different hardware to achieve the best efficiency and performance. AI is a transformative technology that will permeate every aspect of life and be ubiquitous. Qualcomm expects that the market for AI's hardware and software ecosystem will reach $160 billion by 2025. Qualcomm has always stressed that the development of AI is undergoing transformation and is moving from the cloud to the terminal. On the one hand, the computing power of the terminal is improved; on the other hand, the requirements of the user connection experience, privacy, and reliability all promote the transfer of the AI ​​to the terminal side. High-performance, low-power devices required for AI on the terminal side, using wireless communication technology to achieve the connection between the device and the cloud are areas that Qualcomm focuses on. Qualcomm's R&D in machine learning is divided into two parts: the first part, the most efficient machine learning processor for the terminal, and the various use cases that run our partners and partners; the second part will be devoted to researching the terminal side AI. Related use cases, such as security, connection management, photo taking, etc., but also hope that companies in the ecosystem, whether it is Google application developers, or Baidu, Tencent, Alibaba, can use smart terminals to obtain data, develop cloud side artificial intelligence, constantly Improve the cloud training algorithm and put it on the terminal to run. "Our strategy in machine learning is to hope that Xiaolong will become the platform of choice for running machine learning algorithms on the mobile terminal side." Christiano Amon told reporters that in order to achieve this goal, we will focus on terminal-centric technology. R&D and the partner's algorithms are added to the mobile experience. Nowadays, the computing power of mobile chips is already very strong. The PC with Qualcomm's previous generation flagship processor, Xiaolong 835, will be available soon. PCs can enjoy the same experience as mobile phones. In terms of connectivity and endurance, there will be Great improvement, no need to carry the power cord, available all the time, anywhere, keep the power on, rewriting the user experience. In terms of performance, the mobile chip has provided a strong guarantee for the hardware of AI. Taking a different path of AI technology development, in fact, it will not be very different for users. The application aspect is the key factor affecting users. . The core of the artificial smartphone is to constantly learn the various habits of the user, from observing the user to understanding the user, and prejudging the user's behavior in different scenarios, letting the machine understand humans through various human-computer interactions. In this way, everyone's mobile phone will be transformed into a personal assistant, and is unique. At present, the application of AI is more focused on photographing, image recognition and so on. The application of intelligent scenes of AI is still very scarce. Qualcomm sees that there will be explosive growth in the future of AI applications, which will focus on augmented reality, immersive gaming experiences, voice and health monitoring. Christiano Amon acknowledged that although Qualcomm started developing AI technology very early, it did not promote AI too much in the past, mainly because the past AI use cases were not very mature. But now, many AI use cases have begun to become a reality. He believes, "We will see AI use cases ushered in explosive growth in the future." For example, for example, a type of security-like application, the Snapdragon 845 places great emphasis on security features. Machine learning will know the usual behavioral habits of users. If your behavior is different from usual, the terminal will know that there may be security issues. In addition, machine learning will know how you use your phone, and where you are, whether it is good to use Wi-Fi or cellular networks, and how to switch between them; A more important type of application comes from companies that collect vast amounts of data. They will train the algorithms in the cloud and then run them on the Snapdragon platform. He added that although many use cases are beyond our imagination, these use cases will run from the cloud to the Xiaolong platform. A typical example is shopping. For example, Amazon or Alibaba, they will know what similar people will buy and then push ads for them. Mobile phones can also know what most people do in a particular location, compare them with cloud data, and then provide recommendations for you. Qualcomm's goal is to make the Snapdragon platform the best and become the platform of choice for supporting many different use cases. At present, Qualcomm has established cooperation with many partners, including Google, Facebook, Tencent, Baidu, Shangtang Technology, Alibaba, etc., to support the creation of AI-related functions. At the moment, Apple, Samsung, and Huawei have all tested the water and launched some innovative artificial intelligence applications, such as voice assistants and face recognition, which make the mobile phone already have the prototype of artificial intelligence. Apple iPhone X's face ID is added to the artificial intelligence machine learning AI algorithm to support new functions such as Face ID and animation emoji animoji. Siri's human-computer interaction capability is also greatly improved; the newly released glory V10 has "system" "A negative screen" has the function of providing users with the usual life app and one-click direct access; Samsung's Bixby voice assistant function is a typical representative of AI's integration of the app, including voice, visual, homepage, reminder four Large functions, in which these users can freely switch and combine. Whether AI can drive a new round of growth in mobile phones is still unclear. It is undeniable that before the advent of the 5G exchange, the AI ​​gave a bright light to the mobile phone market of “innovation is weak†and “thousands of one sideâ€. With the release of Xiaolong 845, Qualcomm will lead many Android camps in 2018. Mobile phone manufacturers have entered the AI ​​era. The impact of the AI ​​mobile phone on the user experience is subtle. It is difficult to perceive its existence. When one day, you find it inseparable from it, you will be surprised to find that it is already an extension of your body.
Any QI enabled device like iPhone X/ 8/8Plus, Galaxy, Samsung S7,S7 Edge, S6,S6 Edge, Nexus 4/5/6 (NOTE: Other device have no wireless charging function except putting on extra receivers.)
Starts the moment you place down any Qi-enabled device or device equipped with a Qi-compatible cover. No cables or USB interface required. Input: AT LEAST 5V/2A; Output: 5V/1A Wireless Charger,Universal Wireless Charger,Portable Wireless Charger,Mobile Wireless Charger Shenzhen Waweis Technology Co., Ltd. , https://www.waweis.com
