Thread Content
From July 26 to 28, 2025, the World Artificial Intelligence Conference and High-Level Meeting on Global Governance of Artificial Intelligence was held in Shanghai. Under the theme of \"An Intelligent Era: Working Together for a Shared Future,\" the conference showcased the latest advancements in AI technology and industry trends, as well as the newest practices in global governance. The scale of the exhibition and the range of participants in the area of knowledge creation both reached record levels. During the conference, the State-owned Assets Supervision and Administration Commission of the State Council announced the first batch of 40 strategic, high-value applications of artificial intelligence by central state-owned enterprises, covering 16 key sectors such as energy and power, industrial manufacturing, and healthcare; China National Petroleum Corporation’s project on \"research and development of high-performance synthetic rubber materials\" was selected for inclusion in this list. In addition, the Kunlun large model made a notable appearance at the Exhibition of Innovations in China’s Artificial Intelligence Industry held as part of the conference. “The “Research and Development of High-Performance Synthetic Rubber” scenario is built based on the Kunlun Large Model. Synthetic rubber has a complex structure and a lengthy production process; its development relies on trial and error, and there are no mechanism models that can accurately describe the complex structure and multidimensional properties of rubber. As a result, the development costs are high and the iteration cycles are long. “\"Research and development of high-performance synthetic rubber materials\" achieves technological breakthroughs in areas such as rubber molecular structure design, performance prediction, and process optimization by creating high-quality synthetic rubber datasets and training large models for synthetic rubber. Once the application is launched, researchers will be able to flexibly set molecular components and formulation ratios to predict key properties of rubber materials. This is expected to increase the efficiency of rubber molecular structure design and formulation screening by over 20%, contributing to breakthroughs in AI For Science within the energy and chemical industry. The Kunlun Large Model is the first industry-specific large model in China’s energy and chemical sector to be **registered**. To date, it has been trained into a language model with 300 billion parameters, a vision model with 4.4 billion parameters, and a multimodal industry model with 80 billion parameters. Additionally, eight specialized large models have been developed for areas such as seismic processing, seismic interpretation, well logging processing and interpretation, and timing analysis in petrochemical processes. Furthermore, 100 application scenarios within the energy and chemical industry have been created, including intelligent full-waveform inversion of seismic data and optimization of process operations in ethylene production facilities from ethane. An ecosystem conducive to the innovative development of artificial intelligence has thus taken initial shape. At the meeting held on February 19 this year to further advance the \"AI+\" initiative among central state-owned enterprises, the State-owned Assets Supervision and Administration Commission of the State Council launched a special project aimed at developing strategically important and high-value application areas. In April, the State-owned Assets Supervision and Administration Commission of the State Council released a batch of outstanding achievements in the development of high-quality artificial intelligence datasets across more than 10 industries and 30 state-owned enterprises; China National Petroleum Corporation’s \"High-quality Dataset for Oil and Gas Seismic Exploration\" and \"High-quality Dataset for Geophysical Logging of Underground Resources\" were among those selected.