On July 15, the Haikou Municipal Bureau of Science, Industry, and Information Technology announced that Haikou will promote the pilot construction of a national data annotation base, focusing on creating a demonstration base characterized by“AI data internationalization” and “satellite remote sensing and mapping.” During the two-year pilot construction period, Haikou will leverage its geographical advantages to provide value-added services such as “on-demand data processing,” further developing a specialized data annotation industry cluster, establishing international data service capabilities, and participating in the international data element market.
In May of this year, the National Data Bureau announced the first batch of seven national data annotation pilot bases, with Haikou being the only city in the province to be selected. What is data annotation? An official in charge from the Haikou Municipal Bureau of Science, Industry, and Information Technology explained that data annotation refers to the process of converting raw data, such as images, text, voice, and video, into machine-recognizable information through processing. Data annotation is a crucial step in the operation of artificial intelligence algorithms and plays an important role in aggregating data resources, improving data quality, and enhancing the value of data elements.
Currently, the pilot construction of the data annotation base in Haikou primarily relies on the“Baidu Artificial Intelligence Data Annotation Industry Base.” Additionally, Hainan Changguang Satellite Information Technology Co., Ltd. is also engaged in aerospace data annotation services.
On the morning of July 15, at the Haikou Digital Technology Innovation Platform’s Baidu Smart Cloud (Haikou) AI Basic Data Industry Base located in Xiuying District’s China Merchants Tower, AI data annotators were skillfully judging and marking content data generated by large models via a data annotation platform, advancing AI data annotation projects for clients nationwide.
The base, with a total construction area of 11,400 square meters, has attracted over 10 ecosystem partner companies since its operation began in April 2023, with more than 1,000 people employed in data annotation.
Meanwhile, Changguang Satellite is also advancing the construction of its annotation base. Located on the fifth floor of the Haikou MoLi Innovation Workshop, the project is being built in two phases. The first phase, covering 2,447 square meters, is about to be put into use. The second phase is expected to start in 2025 and complete in 2026, forming an industry base with a total area of 5,000 square meters to be established as a high-quality data annotation base.
During the two-year pilot construction period, Haikou plans to focus on six areas: technological innovation, industry empowerment, ecosystem cultivation, standards application, talent employment, and data security. In terms of technological innovation, efforts will concentrate on data governance stages such as data collection, cleaning, annotation, synthesis, and detection, tackling a series of related intelligent and automated data annotation technologies, and developing an integrated intelligent annotation platform. Simultaneously, focusing on data from six industries—cross-border trade, finance, agriculture, cultural tourism, education, and remote sensing and mapping geography—Haikou aims to complete a total annotated data volume of no less than 1,000 TB. The plan includes attracting five leading data annotation companies, cultivating and introducing 50 annotation companies, creating employment for tens of thousands of data annotators, and directly boosting the data annotation industry’s scale to 1 billion yuan.
In the next phase, Haikou will draw on relevant policies from developed regions in nationwide in the data annotation industry, focusing on rent subsidies, production element support, and talent training and introduction subsidies. By the end of September 2024, Haikou plans to introduce support policies for the city’s data annotation industry development. Leveraging the advantages of Baidu’s Wenxin Yiyan large language model and Changguang Satellite’s leading enterprises, Haikou will engage in industrial chain investment promotion, vigorously attract ecosystem partner companies, promote the development of the AI data annotation industry’s upstream and downstream industrial chains, and build a national-level digital economy industry cluster characterized by Haikou’s data-driven, intelligent, and ecologically integrated features.
信息来源：海南日报