#Today’s news#
1. Xu Jiayin: Invest 45 billion in new energy vehicles in the next three years
On November 12, Xu Jiayin, chairman of the board of directors of Evergrande, said that the annual production capacity of Evergrande’s top ten production bases in the world will be 500,000 vehicles, and the initial stage will be 100,000 vehicles. In the next 5-10 years, Evergrande plans to produce 1 million vehicles per year; in 10-15 years, it will reach 5 million vehicles. At the same time, Evergrande will cooperate with Koenigsegg to produce 1,000 supercars per year. In addition, Xu Jiayin introduced 3-point planning. First, ten complete vehicle production bases will be built in China, Sweden and other places, and at the same time supporting core parts and components bases. It is decided to develop 15 models at the same time to achieve full coverage of the product line. He also said that he plans to invest 45 billion yuan in new energy vehicles in the next three years. The specific plan is to invest 20 billion yuan in the first year, 15 billion yuan in the second year, and 10 billion yuan in the third year. The Hengchi brand’s first new energy vehicle “Hengchi 1” will be unveiled in the first half of 2020.
2. 1,200 people in the Apple AR headset detail full exposure project
On November 12, according to foreign media The Information, at an internal meeting of Apple in October this year, Mike Rockwell, Apple’s vice president in charge of AR/VR projects, told at least 1,000 Apple employees Headquarters employees shared Apple’s AR product roadmap, as well as new details about Apple’s AR headsets and features. The core information includes: 1. Apple may release AR headsets in 2022 and AR glasses in 2023. 2. Apple’s AR headset, codenamed N301, looks similar to Oculus Quest and has a mix of AR and VR functions. 3. Apple’s AR glasses, code-named N421, look like a pair of sunglasses, and Apple believes AR glasses can replace iPhone demand in about a decade. 4. As early as 2015, Apple has been exploring AR/VR and building prototypes of AR/VR devices on the T288 project. 5. By the end of 2018, the number of Apple employees working on AR projects across multiple departments is estimated to be around 1,200, with hundreds of them working directly on AR headset projects and dozens on AR glasses projects.
3. The University of Science and Technology of China is the first to realize the simultaneous classification of multiple quantum correlations
On November 12, the team of Academician Guo Guangcan from the University of Science and Technology of China made important progress in the intersection of artificial intelligence and basic research on quantum mechanics. The laboratory Li Chuanfeng, Xu Jinshi and others cooperated with Professor Weng Wenkang of Southern University of Science and Technology and Researcher Ren Changliang of Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences and others to apply machine learning technology to the study of fundamental problems of quantum mechanics. The first experiment realized the algorithm based on machine learning. Simultaneous classification of multiple nonclassical associations. The result was published in the international physics authoritative journal “Physical Review Letters” on November 6, 2019.
#artificial intelligence#
1. Facebook announces the latest CV research: fine-tuning the face
On November 12, Facebook announced the latest research results in the field of computer vision at the ICCV conference, which can enable machine learning models to anonymize faces, improve hand movements, and provide clothing suggestions. . It is reported that its “face-changing” technology is not completely “replaced” with different faces, but fine-tuned on the basis of the original appearance, so as to realize identity concealment and avoid traditional rough “physical occlusion” (such as masks).
2. NVIDIA announces five key priorities for 2020
On November 12, Ian Buck, vice president and general manager of accelerated computing at NVIDIA, recently announced five key areas of focus for NVIDIA in 2020, including software-defined 5G infrastructure, AI accelerated medical care, autonomous vehicles, autonomous machines and AI supercomputers . NVIDIA is enabling new intelligent capabilities for robots through its Jetson AI platform, he said. The platform consists of a series of system-on-module motherboards with CPUs, GPUs and various other high-performance chips. In addition, he said, “We are turning the supercomputers we are making into AI supercomputers.”
3. Ministry of Education: Specialties supplement 9 majors such as hydrogen energy technology AI
According to the information on the official website of the Ministry of Education on November 12, according to the “Administrative Measures for the Setting of Higher Vocational Education (Specialty) Majors in Ordinary Colleges and Universities”, on the basis of supplementary professional suggestions submitted by relevant schools and industries, the Ministry of Education organized research to determine the 2019 year. Nine new majors, including hydrogen energy technology and artificial intelligence, are added, which are hereby announced and will be implemented from 2020.
4. Twitter releases its first draft anti-deepfake strategy
News on November 12th, yesterday, Twitter officially released its first draft anti-deepfake strategy. The company said it would remove deepfake content from Twitter if it threatened someone’s physical safety or could cause serious harm. Currently, Twitter is still soliciting public comments for the draft.
5. Google’s new architecture: it only takes 37ms for a single CPU to process 1s video
On November 12, the Google robotics team proposed a general approach to designing a family of tiny neural networks for video understanding tasks. These micro-neural networks can achieve fairly high accuracy and run efficiently, either in real-time or at faster speeds. They take 37-100ms to process a video clip of about 1 second on a CPU and 10ms on a GPU, which is 100 times faster than current models. The researchers refer to these networks as Tiny Video Networks (TVNs) because such networks require only a minimal runtime, a first for a video model.
6. Abu Dhabi Capital Group and UBTECH signed a 2.5 billion cooperation
It was reported on November 12 that on November 11, local time, Ubisoft signed an AI education project cooperation agreement with the UAE Royal Strategic Partner, a subsidiary of Abu Dhabi Capital Group. The two sides jointly launched an artificial intelligence talent training program. The amount of this cooperation project reached 362.4 million US dollars, equivalent to about 2.5 billion yuan. It will be implemented in four years and will build artificial intelligence teaching laboratories for a total of 1,310 primary and secondary schools in seven emirates of the United Arab Emirates.
7. The emerging AI technology federated learning standard may be introduced next year
On November 12, it was learned from the fourth meeting of the IEEE P3652.1 (Institute of Electrical and Electronics Engineers Federated Learning Infrastructure and Applications) standard working group that the draft federated learning standard is expected to be launched in February 2020, and the official standard is expected to It will be launched in the first half of 2020. At present, federated learning, an emerging AI technology, has been applied in many fields such as finance, medical care, and urban management. The IEEE Federated Learning International Standard Project was initiated by WeBank and approved in December last year. It is the first project in the world to set standards for the artificial intelligence collaborative technology framework.
8. Google is rumored to be secretly collecting health data on millions of Americans
On November 12, it was reported that Google’s “Project Nightingale” secretly collected the personal health data of millions of Americans without the knowledge of patients and doctors. The data collected includes laboratory results, doctor’s diagnoses and hospitalization records, as well as the patient’s complete health history and other information such as name and date of birth. Google, which launched the project last year in partnership with U.S. medical group Ascension, is using the data to design software that uses machine learning to personalize care for different patients.
#AIchip#
1. Loongson Micro Semiconductor project settled in Xiangdong District, Jiangxi
It was reported on November 12 that the Foster Smart Manufacturing Industrial Park project and the Loongson Micro Semiconductor project officially settled in Xiangdong District, Pingxiang, Jiangxi, with a total investment of 6 billion yuan. It is understood that the project focuses on the development of integrated circuits and microelectronics, and integrates upstream and downstream enterprises to create five major sectors including semiconductor chips, satellite navigation, intelligent manufacturing, industry-university-research and development institutes, and 5G mobile communications. The first phase of the intelligent manufacturing industrial park invested 3.5 billion yuan in the Loongson Micro-semiconductor project, which mainly produces the packaging and testing of DFN, QFN, SOP, DIP and other series of products.
2. Qualcomm will release the Snapdragon 865 using the Arm A77 architecture
On November 12, according to foreign media GSMArena, Qualcomm will hold Snapdragon Tech Summit 2019 in Hawaii, USA from December 3 to December 5. Qualcomm is said to be launching its next flagship chipset, the Snapdragon 865, at the summit. According to foreign media, the CPU performance of the Snapdragon 865 is about 20% higher than that of the Snapdragon 855. Some Korean media said that Qualcomm has selected Samsung to manufacture the Snapdragon 865 chipset. Relevant information shows that the Snapdragon 865 chipset includes one Cortex A77 at 2.84GHz, three Cortex A77 at 2.42GHz and four A55 cores with a main frequency of 1.8GHz.
3. It is reported that Nvidia will push the ampere architecture in March next year
On November 12, according to reports, Raymond James analyst Chris Castle said through HKEPC that Nvidia delayed the launch of Ampere by several months. The Ampere Architecture will be released at GTC at the end of March next year, and the Ampere Architecture will be used first in a new AI product for data centers. The GeForce RTX 3000 series graphics cards based on the Ampere architecture will be launched in a few months, and the RTX 3080 may debut at Computex 2020.
4. Xilinx unified software platform Vitis is open for download
On November 12, Xilinx Vitis unified software platform and optimized open source library officially opened for free download from now on. According to reports, the Vitis platform does not mandate a proprietary development environment, but can plug into common software development tools and utilize open source libraries optimized for Xilinx hardware. Xilinx also has a developer website that allows developers to easily access examples, tutorials, and documentation.
The Links: 3HAC037017-001 3HAC047184-003
