Five large models in Beijing are the first in the country to complete online registration
Large models are becoming a new hot spot in the field of artificial intelligence.
Recently, Xu Yali, deputy director of the Cyberspace Administration of Beijing Municipal Committee and director of the Beijing Branch of the National Center for Artificial Intelligence, revealed at the "2023 Beijing Artificial Intelligence Large Model Scenario Integration and Industrial Development" event that as of now, Beijing has 5 large models in the country. It was the first to complete the online registration, and more than 20 other companies submitted registration materials according to the "Interim Measures for the Management of Generative Artificial Intelligence Services" issued by seven ministries and commissions including the Central Cyberspace Administration of China.
Liu Chenglin, deputy director and researcher of the Institute of Automation of the Chinese Academy of Sciences, said that in recent years, deep learning methods have driven the rapid development of specialized intelligent technologies and have been successfully applied in fields such as visual monitoring, speech recognition, intelligent robots, autonomous driving, and scientific research. In the future, with the development of large model technology, general artificial intelligence, brain-like intelligence and other aspects will continue to develop and continue to penetrate into all levels of society.
At present, scenario-driven technology has become an important model for large-scale technological innovation and is expected to form new industrial formats and economic growth points. In order to promote the implementation of relevant projects, the Beijing Municipal Bureau of Economy and Information Technology has actively completed implementation promotion work in various aspects such as policy support, project support and technology promotion. "On the one hand, we will improve Beijing's computing power supply capacity in multiple dimensions, vigorously promote the construction of public computing power centers, and actively guide the transformation and upgrading of old small and scattered data centers into intelligent computing power centers. On the other hand, we will actively promote the construction of data pioneer areas and ensure safe and secure computing power centers." Under the premise of providing qualified enterprises with high-quality data resources for the implementation of large-scale model training, Wang Lei, deputy director of the Beijing Municipal Bureau of Economics and Information Technology, said that Beijing will also accelerate the implementation of large-scale model applications and research and compile to promote the implementation of large-scale model training. Relevant solutions for the implementation of large model applications.
In terms of scene integration, Feng Tianyi, deputy director of the Digital Health Department of the Cloud Institute of China Academy of Information and Communications Technology, introduced that the current large AI model has been used in seven major fields: medical services, patient services, medical science, teaching and research, public health, traditional medicine, drug supply, and operation management. Some exploratory applications. However, while large models demonstrate potential and value in applications, they also face risks in terms of technological maturity, medical data quality and security, and medical quality and fairness. Therefore, Feng Tianyi suggested that promoting development and regulatory standards should go hand in hand. Establishing and improving the large-scale model standard system and scientific and technological ethics governance system in the medical and health industry is the key to promoting the high-quality development of the large-scale model industry in the medical and health industry.
In another vertical scenario, Huang Lei, director and vice president of the research institute of Hanwang Technology Co., Ltd., said that the combination of large models and scenarios can enhance the basic capabilities of artificial intelligence technology. For example, the Hanwang Tiandi large model has nine basic capabilities: multi-modality, semantic understanding, logical reasoning, mathematical calculation, knowledge search, tool invocation, content generation, multi-language and multi-round dialogue. In practical applications, it shows data privatization, It has five major characteristics: low-cost computing power, deep specialization, real-time knowledge, and accurate generation.
Wu Zhengwu, deputy chief engineer of the China Judicial Big Data Research Institute, said that the large vertical model of the legal service industry created by the China Legal Research Institute is being applied in scenarios such as judges' case handling, legal knowledge Q&A, lawyer services, legal consultation, corporate compliance review, and laws and regulations review. play an active role.
The white paper "Beijing City of Large Models—Industrial Map and Innovation Ecology" shows that the scale of my country's artificial intelligence large model industry is booming. The white paper suggests that if my country's large model industry continues to reach a new level, it needs to further improve the relevant policy system, emphasize original technological innovation, expand application scenarios, and cultivate industrial "soil."
