Shandong province unveiled seven vertical AI models, collectively called the Qilu Seven Stars, at a three-day AI industry conference in Jinan. The models are intended as industry-specific solutions rather than general-purpose foundation models, and the conference included a main forum, more than 20 thematic sub-forums and a series of special activities to match supply and demand, connect industries and foster collaborative innovation in AI.
Vertical AI models tailored to Shandong industries
The seven models were presented as specialist champions rooted in Shandong’s industrial strengths. A highlighted example is a mining large model launched jointly by Shandong Energy Group, Yunding Technology and Huawei. At the main transport system of the Lilou coal mine, miners previously walked nearly 10 kilometers a day to inspect equipment; AI cameras deployed with the model can now remotely detect anomalies such as large coal lumps and anchor bolts with 98 percent accuracy.
Building a vertical large model for complex underground conditions requires core industry data largely sourced from real mining sites. Shandong Energy Group owns nearly 80 coal mines that cover almost all geological conditions found in Chinese coal mines. Millions of underground images and hundreds of billions of pieces of production data have been converted into high-quality industry datasets used to train and test such models.
Li Husheng of the China Industrial Control Systems Cyber Emergency Response Team said, “Shandong’s core advantage is its complete industrial categories and many application scenarios, which provide fertile ground for vertical large models to grow.” The province’s comprehensive industrial system and accumulated industrial process data are presented as unique advantages for moving AI from research to large-scale application.
Scope and capabilities of the Qilu Seven Stars
The seven models cover chemical safety, light manufacturing, underground mining, port scheduling, recycling, metallurgy and precision manufacturing. They include a ‘smart eye’ for chemical production safety, a ‘super brain’ for flexible light manufacturing and a ‘digital guardian’ for frontline underground operations. Other capabilities described include scheduling berth plans for 10,000-ton vessels with minute-level precision, supporting intelligent sorting in recycling, offering an optimal solution for energy saving and carbon reduction in metallurgical blast furnaces, and helping achieve extreme precision in core equipment for precision manufacturing.
