Xing Huina:
Due to time constraints, we will take the last two questions.
Economic Information Daily:
We are concerned about central SOEs' development in the artificial intelligence industry. Previously, the SASAC has launched an initiative to advance AI-driven industrial upgrades. What progress has been made so far? What's the follow-up plan? Thank you.
Wang Hongzhi:
Thank you for your questions and your attention to central SOEs' development in the artificial intelligence industry. Artificial Intelligence is an important driving force for the new round of scientific and technological revolution and industrial transformation. It is also a key engine for the cultivation of new quality productive forces. For a period of time, central SOEs have made active explorations and bold attempts, with positive progress being made in the fields of computing power, data, large models and scenario applications. The development of artificial intelligence relies on cooperation and ecosystems. We will work with all sectors of society to make joint efforts within infrastructure, supply, demand and policy in order to advance industrial development with higher quality and faster speed, better empowering the high-quality development of the economy and society. With regard to this, Mr. Fang Lei will give you a detailed introduction.
Fang Lei:
Well, let me answer your questions specifically. In recent years, the SASAC has been thoroughly implementing General Secretary Xi Jinping's important expositions on accelerating the development of artificial intelligence. We have conformed with the situation and taken the initiative to accelerate the development of artificial intelligence, stressing overall planning in our work. We have also taken it as the major direction for initiatives to advance industrial upgrading and foster new enterprises for emerging industries as well as formulated and issued action plans. We have organized symposiums on advancing the development of artificial intelligence for central SOEs, launched the AI+ special program, and, as Mr. Wang Hongzhi just mentioned, have made simultaneous efforts in computing power, data, algorithm and applications, with positive progress achieved mainly in four aspects.
The computing power infrastructure has been consolidated at a faster pace. As of the end of June, the scale of intelligent computing power of central SOEs had doubled year on year. Massive computing clusters comprising 10,000 GPUs have been built in cities including Shanghai and Hohhot. The computing power platform has initially realized multi-core heterogeneous computing power scheduling, effectively supported the training and iteration of general AI large models with a parameter of 100 billion and above, and become more independent and controllable, providing solid support for large-scale application.
The development of AI basic models has been advanced steadily. We have supported general AI large models, such as Jiutian and Xingchen, to strengthen technical breakthroughs and ecological cultivation, and to push forward multimodal large models and cross-modal image generation capacities to catch up with the leading level in China. We have continued to promote the empowerment of large models to industries and have led central SOEs, private enterprises and institutions of higher learning to jointly build models for the industry.
Scenarios for application have been built at a faster pace. Over a hundred application scenarios have been built in areas such as new materials, communications and manufacturing. Large AI models have been primarily applied in areas including electricity, transportation and petrochemicals. Significant progress has been made in reducing costs and increasing efficiency in the areas of scientific research, production and customer services. A batch of high-quality datasets have been developed in such fields as logistics, finance and healthcare. Data governance tools have been developed and value-added services have been gradually offered, such as data processing.
The role in driving the development of industry has been continuously enhanced. Many central SOEs released standard products including software factories and 5G+AI new calling services at the 7th Digital China Summit and the World Artificial Intelligence Conference. They have strengthened cooperation with private enterprises such as Huawei and iFlytek in areas including intelligent networked vehicles and energy and chemical industries, formed supply and demand docking and have jointly developed and promoted the application of large AI models for the industry, creating a sound industrial ecosystem.
Next, the State-owned Assets Supervision and Administration Commission of the State Council (SASAC) will accelerate the development of the artificial intelligence industry through application demonstrations. Centrally administered state-owned enterprises (SOEs) are confident, determined and capable of seizing strategic opportunities. They will work with various partners to provide intelligent computing resources, transform application scenarios and cultivate the industrial ecosystem. This will involve first to thoroughly explore and fully open high-value scenarios, as well as conducting supply-demand cooperation to create a set of industry application demonstration benchmarks. Second, exploring high-quality dataset standards and construction paths, thus building high-quality datasets for key industries in batches. Third, orderly advancing the construction of intelligent computing centers and platforms for both scheduling and operating computing power, as well as strengthening intelligent computing capacity supply to better serve small and medium-sized enterprises. Fourth, focusing on cutting-edge technologies to enhance foundational large model capabilities, improving large model evaluation systems, exploring and establishing industrial development communities and stepping up efforts to empower others. Thank you!
