The development of the artificial intelligence industry has been heated up, and the formulation of relevant standards has become even more urgent.
On January 18, the National Standardization Management Committee formally established the National Artificial Intelligence Standardization Group and the Expert Advisory Group. The head of the overall group is Zhao Bo, Dean of the China Electronics Standardization Institute, and the deputy head candidate is related to Alibaba Cloud, HKUST, China Academy of Sciences, Peking University, Beijing University of Aeronautics and Astronautics, and Institute of Instrumental Technology and Economy Tubes, experts. Pan Yunhe, an academician of the Chinese Academy of Engineering, heads the expert advisory group.
It is understood that the National AI Standardization Group assumes the role of overall coordination, planning and layout in the artificial intelligence standardization, and is responsible for carrying out international and domestic standardization work on artificial intelligence, including drawing up China's artificial intelligence standardization plans, systems and policies, and coordinating relevant national standards and technical content. Technical responsibility, the establishment of AI basic performance standards and industry application standards, such as the transmission mechanism. The expert consultation group is composed of well-known experts and scholars in the field of artificial intelligence in the country, and is responsible for providing advice to the general group on China's artificial intelligence standardization planning, systems, policy measures, and other aspects.
“Standardization work has a fundamental, supporting, and leading role in the development of artificial intelligence and its industries. It is both the key point for promoting the development of industrial innovation and the commanding height of industrial competition.” Zhao Bo said, “Although China is in the field of artificial intelligence It has a good foundation, but the infrastructure, policies, regulations, and standards that are suited to the development of artificial intelligence are still in urgent need of improvement."
Standard Leading Industry
The establishment of the National AI Standardization Overall Group and the Expert Advisory Group is a response to the issuance of the “Three-year Action Plan for the Development of a New Generation of Artificial Intelligence Industry (2018-2020)” (hereinafter referred to as the “Action Plan”) by the Ministry of Industry and Information Technology last month.
The previous "Action Plan" has made specific plans for the development of artificial intelligence in China for the next three years. One of the four major development goals involves the establishment of an artificial intelligence industry support system and preliminary establishment of an artificial intelligence standard system, testing and evaluation system, and Security system framework.
“The advancement and improvement of artificial intelligence standards are related to the healthy development of the industry and the strength of the product's international market competitiveness,” Zhao Bo pointed out.
In fact, developed countries including the United States, the European Union, and Japan have all attached great importance to artificial intelligence standardization. The 21st Century Business Herald reporter learned that both the “National Artificial Intelligence Research and Development Strategy Plan” issued by the United States, the “Human Brain Project” issued by the European Union, or the “Artificial Intelligence/Big Data/Internet of Things/Network” implemented by Japan "Safety Integrated Projects" all put forward reinforcement deployments around core technologies, top talents, and standards and regulations, in an effort to seize a new round of scientific and technological dominance.
According to Zhao Bo, large-scale, commercially available products and applications have appeared on the market. However, it is necessary to solidify the technical achievements with standardized means to achieve rapid innovation and promotion; such as face recognition systems, smart speakers, service robots and other product quality Uneven, but also need to unify standards to improve the quality of products and services.
However, around the ethical issues such as the "difficulty of trams" in the field of automatic driving and the privacy of users of fingerprints leaked from iPhones, it is also necessary to formulate relevant safety standards and norms to ensure that information security and intelligent systems serve human ethics. In addition, under the current situation in which industry giants create ecosystems such as self-owned deep learning frameworks by using open source algorithms and platform interface bindings, user data information is more difficult to migrate, and data islands are formed, making it easier to achieve standardization among vendors. Interoperability and collaboration.
“Either basic standards, technical standards or ethical standards are the key considerations for the future national AI standardization group.” Wu Wenjun, deputy head of the National AI Standardization Group and professor at Beijing University of Aeronautics and Astronautics, told the 21st Century Business Herald reporter, “For Research and discussion in these areas will go hand in hand.
Standard architecture
By analyzing the development status of artificial intelligence at home and abroad, we can find where AI standardization needs are.
“As an enterprise in the artificial intelligence industry, we can feel that everyone has an urgent need for standardization.” Jun Jun, Vice President of HKUST, said: “The first thing that comes to mind is the system framework, followed by the capabilities, and the combination of artificial intelligence technology and the industry should have What capabilities can be defined as wisdom, which needs further clarification. In addition, artificial intelligence interfaces, data standards and evaluation are also very important directions for industry standardization.”
Around these requirements, artificial intelligence standard architecture has also formed. “The artificial intelligence standard architecture includes six parts: basic standards, platform/support standards, key technical standards, product and service standards, application standards, and safety/ethical standards,” said Zhao Bo.
The Artificial Intelligence Standardization White Paper released by the China Electronics Standardization Institute specifically describes the architecture. Among them, the basic standards cover four categories: terminology, reference architecture, data, and test evaluation. They are located at the lowest level of the AI standard architecture. The platform/support standard is the integrated integration of artificial intelligence hardware, software, network, and data, and plays a role in the artificial intelligence standard architecture. The key technical standards mainly focus on natural language processing, human-computer interaction, computer vision, biometrics recognition, and VR/AR, and provide support for the practical application of artificial intelligence.
The application standards are located at the top level of the artificial intelligence standard architecture and face the specific needs of the industry. The product and service standards involve the related standards of intelligent products and new service models formed in the field of artificial intelligence technologies. Safety/ethics standards run through other parts and support the entire industry.
"Huawei initially believed that artificial intelligence is an enterprise's enabling technology. By developing artificial intelligence technology, it can help companies achieve internal efficiency improvement." Song Zhe, vice president of cloud strategy and business development at Huawei, told a 21st Century Business Herald reporter. Focusing on the formulation of current AI standards, taking into account that open source has become a de facto standard in the current industry, Song Zhe-hsuan recommended that open source be introduced as an important factor in the entire standard-setting process so that real standards can be launched well after launch. practice.
Transparent Cards,Translucent PVC Card,Transparent Clear Magnetic Card
Smart Cards,RFID Cards And Tag Co., Ltd. , http://www.cnsmartcards.com
没有评论:
发表评论