我省出台制造业领域数据治理参考指引破解数据“采不准、格式乱”难题

Core Insights - The newly released "Guidelines for Data Governance in the Manufacturing Sector Facing Artificial Intelligence (2026 Edition)" aims to guide manufacturing enterprises in Jiangsu Province to systematically conduct data governance and effectively utilize data governance technologies and methods for artificial intelligence [1] Group 1 - The core of artificial intelligence applications relies on high-quality data for model training, inference, and iteration, making data governance essential for ensuring data quality [1] - The deepening of artificial intelligence applications is driving a shift in data governance from "passive compliance" to "proactive value-driven" approaches [1] - Current challenges in the manufacturing sector include data "silos" and "distortion," lack of data governance and standardization, and disconnection between data and application scenarios, which severely restrict the supply of high-quality, scenario-based datasets [1] Group 2 - The guidelines categorize data governance into three levels: entry-level, basic, and advanced, tailored for enterprises of different sizes and capabilities, providing reference and deployment solutions for AI applications in typical scenarios [1] - The guidelines focus on six core processes: data collection, preprocessing, feature engineering, data labeling, data partitioning, and data augmentation, offering categorized governance paths [2] - Manufacturing enterprises can select appropriate data governance techniques based on their technical foundation, resource conditions, and specific business pain points to maximize data value [2]