iphone x is about to be officially sale, with Apple selling 5.5 million iphone x in 5 hours on the mainland market, with data from multiple dimensions breaking history. After the iphone x sale, someone throws a question, which Android will first learn from the iphone x's ' Face-id '? Domestic mobile phones used to stare at Apple's latest product, and what new features were validated, quickly implemented on their own products, such as touch-id, cancelling headphone holes, three-stage design ... This year, however, the situation may become very different, ' face-id ' will be very difficult to replicate.
Why can't a smartphone replicate Face-id quickly?
Unlike past Touch-id, 3d-touch or canceling headphone holes, Face-id is not just a simple hardware design or integration of a module, its essence is an AI service--based on intelligent face recognition of AI services, the most important is the sound recognition function, But at the same time also Yan stretched out AR expression and other interesting functions. It is based on the chip (A11 bionic processor-driven AI capabilities), end-side sensors (3D depth camera data acquisition capabilities can be compatible with multiple scenarios), Cloud algorithm (face recognition algorithm) of the comprehensive implementation of multiple capabilities, as well as Apple built the certification platform of the blessing ( Only in this way can the FaceId be used for payment and other scenarios), indispensable.
Because of this, the AI function such as Face-id is very difficult to be copied by the general manufacturers, after all, not several handset manufacturers at the same time have three of capabilities. In the anti-Android camp, most vendors simply strive for the start of a high-end processor, only a theme-level UI customization based on Android, or a symbolic cloud service without continuous operation, can not form the true meaning of the core + end + cloud strength. At least in the case of AI, whether it's from Siri's great strides over the years or the success stories of Amazon Echo, it shows that the real intelligence experience can be realized only with hardware and software combined with continuous Internet service. More and more mobile phone manufacturers realize the importance of the core + end + cloud, especially those interested in the card position AI manufacturers, are strengthening these three of capabilities. Has started the ' AI first ' strategy of Google in this year's acquisition of HTC Hardware department to make up for the hardware of the short board, Huawei and Apple is coincidentally in the chip level to introduce the exclusive AI module, among the forefront of the industry, so that the smart phone terminal level of AI computing power to get qualitative upgrade. Why AI want core + END + Cloud Armed Forces combat?
IPhone x A large number of applications AI technology, however, it is not the most AI elements of the smartphone, with the same time released Huawei Mate10, introduced more AI features, smart photography, intelligent power-saving mode, AI accompanying translation functions, System Intelligence optimization, AI personal assistant function ... All this AI function is inseparable from the chip, the system and the integrated support of the cloud. The reason that so much of the AI function, and Huawei Mate10 carrying Kylin 970 chip inseparable, Kirin 970 carried a NPU (neural network unit) dedicated hardware processing unit, specifically for the AI algorithm to provide chip support, processing AI tasks can be 50 times times energy efficiency and 25 times times performance improvement, In order to achieve better image recognition, voice interaction and intelligent photography, such as AI applications.
Powerful chips also need to be used to release their performance, which is the work of the mobile operating system EMUI8.0, in the mobile phone hardware (such as cameras, sensors, screens, etc.), applications, services and chips between the bypass, so that they can fully utilize the ability of the chip, mainly embodied in three aspects: first, Ensure the basic operational efficiency of smartphones. The most intuitive feeling of the user is the mobile phone. EMUI8. 0 through the artificial intelligence technology means carries on the intelligent behavior forecast, the dynamic memory compression, the real-time resources cleanup and so on operation, the reasonable allocation limited handset memory resources, compared with earlier version, the system response time increases 60%, the operation fluency Enhancement 50%, realizes the ' natural fast, the life fast ' ability, It solves the problem that the Android system uses more and more cards. Second, the system level provides a large number of AI services. EMUI8.0 uses AI to provide services at the system level, in addition to providing AI personal assistant, voice helper AI small E, AI dual-fingered screen, such as AI Assistant services, but also intelligent perception of the scene to provide personalized functional services, in the AI eye to know the technology to automatically identify the scene, real-time adjustment parameters, Improve the photo effect. In the future, music, file management, telephone harassment and other basic functions are expected to be upgraded based on AI technology. Third, the system becomes an open window for AI capabilities. Huawei EMUI8.0 not only through the AI technology for the user Service, but also in collaboration with Kirin 970 to open the AI processing capacity to third-party applications, the mainstream AI platform for the industry such as Caffe, TensorFlow acceleration, innovation third-party AI application experience. In this way, all third-party AI applications on Huawei mobile phones can also have better AI processing capabilities, such as Microsoft Translation algorithm application to mate 10 has a better performance. AI-Customized chips + open operating systems are not enough, cloud is also the future of smartphones are inseparable part of the cloud of large data is the basis of AI algorithm evolution, cloud service integration also determines the experience of mobile phones, such as the integration of mobile payment platform, the integration of digital entertainment services, the security of the cloud, storage capacity , intelligent service ability and so on integration; Like EMUI8. 0 for domestic users to create a focus on intelligent function AI personal assistant, voice helper AI small E, AI dual-fingered screen, etc. is based on the cloud service and implementation.
AI application is the Cloud + end, the combination of hardware and software in the industry has been a consensus, Siri can be more and more smart, not without the support of the structure of the cloud services. However, from the iphone x and Huawei mate 10 AI function, the future is only cloud + end is not enough, through the AI chip to strengthen the end of the AI processing capacity in the future will become increasingly important. It can be said that the cloud is the protection of the Air Force, the end is the direct battle of the Army, the core is the future of the Navy, smart phone to do AI, the sea and air armed forces are indispensable. And can foresee, not only is the AI function, more and more smartphone function, all must be based on cloud + END + core synthesis realizes.
Will Apple's Huawei set off a core-cloud arms race? On the eve of World War II, most countries have only navy and army, but today, the smaller countries will build their own maritime and air armed forces. Similarly, smartphones now have only a handful of vendors, such as Apple and Huawei, which form the triple synergy of chips, terminals and clouds, and most of the handset makers only have cloud and end two armies, or even only the army. AI is becoming more and more important in the smart phone, AI decided to experience the smartphone today, more and more mobile phones will find ways to build their own sea and air armed forces.
Huawei and Apple are the leaders of smartphones, and are the first to introduce AI modules at the chip level to strengthen the end-to-end AI capabilities of handset manufacturers. I think Huawei and Apple will no longer be a simple ' feature reproduction ', after all, today's smartphone cannot be won by a single design or function, and the truly competitive features will not be copied at 1:30. As a result, more handset manufacturers will be involved in the chip link, and on the chip to add a special AI module, strengthen the AI processing capacity.
Now, Millet has been involved in the chip link, the introduction of the Pine cone processor small test; Google is likely to be at the chip level after acquiring the HTC hardware team; Samsung processor strength itself is very strong, it has invested in China's artificial intelligence technology chip manufacturers Deep Technology (Deephi Tech), the British AI technology company, including Graphcore, a number of AI companies, Korean media reported that Samsung will be next generation processor Exynos 9810 to join the Kirin 970 similar to the NPU, the next generation of flagship S9 is expected to apply this AI core. At the chip level AI, the formation of the core + end + cloud of the naval and air combat capabilities, is becoming a significant trend, Huawei, Apple is in the mobile phone industry set off a new arms race-no longer a patchwork of configuration and parameters, but to fight the "core-end Cloud" armed forces combat capabilities. Of course, not every mobile phone company has the strength to ' self-reliance ', like many small and medium-sized countries now rely on advanced defense equipment procurement, the future small and medium-sized smartphone manufacturers to build ' navy and Air Force ' also rely on procurement. Qualcomm and other old chip manufacturers are strengthening the chip's AI processing capabilities, Samsung can open the Exynos processor to the Phantom clan, Huawei can do so in the future. However, by purchasing third-party AI chips and technology, it is difficult to achieve deep integration to form a consistent and perfect experience, and it is likely to be by: Huawei, Samsung will be the first to use the formation of their own advantages before the open technology to the third party, and Qualcomm once mastered the core AI technology will let smartphones break the first race, This feeling is not very good for smartphone makers. Therefore, the smartphone manufacturers should take the ' self-reliance ' way, Huawei, Samsung, Apple today has no worries, it is the self-reliance of many years ago, Jobs decided to develop a series of processors, Huawei decided to do Kirin processor, many years ago is also a difficult choice, because they have many years of algorithms, data, patents, Chip, cloud service, system and other ' core + end + cloud ' integration, only today's thick product thin hair.