The reporter learned from the National Data Administration that it has guided seven cities including Hefei and Chengdu to build data annotation bases and explore the development experience of the data annotation industry. As of the first half of this year, the seven data annotation bases have built 524 datasets and served 163 large models. A high-quality dataset is a set of data that has been collected, processed, and can be directly used to develop and train artificial intelligence models, effectively improving model performance. Yu Ying, Deputy Director of the National Data Administration, stated that the National Data Administration is comprehensively accelerating the construction and application of high-quality datasets, and organizing special actions for ecological cultivation. Carry out the collection and demonstration promotion of typical cases of high-quality datasets, and explore benchmark practices in key fields such as healthcare, industry, and transportation; Regularly hold technical exchange activities and conduct in-depth discussions on methodologies for data annotation, synthesis, and high-quality dataset construction; Build a normalized supply and demand docking platform to promote precise matching between data suppliers, technology providers, and scenario application providers. Next, the National Data Administration will focus on building a closed-loop ecosystem of "data annotation+high-quality datasets+models+application scenarios+market-oriented value", promoting the market-oriented allocation reform of data elements and the "artificial intelligence+" action in the same direction and resonance. (New Society)