news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

At a press conference held by the State Council Information Office on July 26, Fang Lei, head of the Science and Technology Innovation Bureau of the State-owned Assets Supervision and Administration Commission of the State Council, said that in recent years, the State-owned Assets Supervision and Administration Commission of the State Council has conscientiously implemented General Secretary Xi Jinping’s important expositions on accelerating the development of artificial intelligence, actively adapted to the situation, and taken the initiative to take action. It has put the accelerated development of artificial intelligence in the overall work of state-owned central enterprises and made it a major direction for the deployment of the Industrial Renewal Action and the Start-up Action. It has formulated and issued an action plan, organized a special meeting to promote artificial intelligence for central enterprises, and launched the "AI+" special action, making simultaneous efforts in computing power, algorithms, data, and applications, and has made positive progress in four major areas.

The computing power base is being consolidated at an accelerated pace. By the end of June, the scale of intelligent computing power of central enterprises had doubled year-on-year, and 10,000-card clusters had been built in Shanghai, Hohhot and other places. The computing power platform has initially realized the scheduling of multi-heterogeneous computing power, effectively supporting the training iteration of general large models at or above 100 billion levels, and the autonomous and controllable capabilities have been gradually improved, providing solid support for the implementation of large-scale applications.

The basic model is being steadily built. Support the "Jiutian" and "Xingchen" general large models to strengthen technical research and ecological cultivation, and promote the multimodal large model and cross-modal image generation capabilities to catch up with the domestic first-line level. Continue to promote the open empowerment of large models, and drive the central enterprises, private enterprises, universities and research institutes in the industry to jointly build industry models.

Application scenarios are being built faster. More than 100 application scenarios have been created in the fields of new materials, communications, manufacturing, etc. Large-scale model applications have been initially realized in the fields of electricity, transportation, petrochemicals, etc., and cost reduction and efficiency improvement have been significantly achieved in scientific research, production, customer service, etc. A number of high-quality data sets have been initially built in the fields of logistics, finance, medical care, etc., data governance tools have been developed, and value-added services such as data processing have been gradually provided to the outside world.

Industrial drive continues to strengthen. Many central enterprises released standard products such as software factories and 5G+AI new calls at the 7th Digital China Summit and the World Artificial Intelligence Conference, and strengthened cooperation with private enterprises such as Huawei and iFlytek in the fields of intelligent connected vehicles, energy and chemical engineering, forming supply and demand docking, jointly creating industry big models and promoting their implementation, and building a good industrial ecology.

Fang Lei said that in the next step, the State-owned Assets Supervision and Administration Commission of the State Council will accelerate the development of the artificial intelligence industry driven by application demonstrations. Central enterprises also have the confidence, determination and ability to seize strategic opportunities, join hands with all parties, and accelerate their development into suppliers of intelligent computing resources, transformers of application scenarios and cultivators of industrial ecology. First, we will deeply explore high-value scenarios and fully open them up, carry out supply and demand collaboration, and form a number of industry application demonstration benchmarks. Second, we will explore the standard system and construction path of high-quality data sets, and build high-quality data sets for key industries in batches. Third, we will orderly promote the construction of intelligent computing centers and computing power scheduling and operation platforms, strengthen the supply of intelligent computing capabilities, and better serve small and medium-sized enterprises. Fourth, we will keep a close eye on the frontier to enhance the basic large model capabilities, improve the large model evaluation system, explore the establishment of a number of industrial development communities, and accelerate the promotion of external empowerment.