High-quality dataset construction and standardization symposium for central enterprises held in Beijing.

In response to the current issues of lack of standardization, varying quality and inefficiency in the construction of high-quality datasets, the National Standards Committee is accelerating the development of technical documents and series standards for high-quality datasets, clarifying the construction path, format specifications, classification requirements, quality evaluation systems, etc. of high-quality datasets, promoting the transformation of artificial intelligence models from usable to better use, and effectively promoting the digital transformation of various industries. Next steps: First, accelerate the development of key standards, focus on the common needs of industries, promote the development of guidelines, format requirements, classification guidelines, quality evaluation specifications, data labeling, data synthesis requirements and other key domain standards for the construction of high-quality datasets. Second, deepen the standard validation pilot, focusing on central enterprises to carry out pilot projects, improve the standard content and create demonstration cases, accelerate the application of more industries. Third, improve the standard pilot platform tools, enhance the practicality and operability of the standards, improve the efficiency of standard validation pilots, and systematically promote the construction of high-quality datasets. Fourth, strengthen international standard coordination, participate in and actively undertake ISO/IEC JTC1/SC42 WG2 international standard work.