en.Wedoany.com Reported - On May 6, 2026, the General Office of the Ministry of Industry and Information Technology and the Department of General Affairs of the National Data Administration officially issued the "Notice on the Joint Implementation of the 2026 'Model-Data Resonance' Initiative" (MIIT Joint Science and Technology Letter [2026] No. 193). Targeting 20 industry sectors including steel, petrochemicals, automobiles, aerospace, electronic components, and information and communications, the notice systematically deploys seven key tasks for the synergistic interaction between artificial intelligence models and data resources. The notice requires each provincial-level region to select at least three key industries, and each central enterprise to select at least one industry to participate in the initiative, aiming to fundamentally establish a virtuous cycle of "data-model-scenario application" mutual reinforcement by the end of 2026.

The core logic of this initiative lies in resolving the structural contradictions of "scattered data, weak models, and disconnected scenarios" currently prevalent in the industrial AI field. The notice explicitly proposes the concept of "Model-Data Resonance," which involves building large-scale, high-quality industry datasets to provide a reusable knowledge base for training industry models. Simultaneously, model evaluation results are used to inversely diagnose quality gaps in the datasets, forming a positive closed loop of "evaluation diagnosis—targeted dataset optimization—model capability enhancement." This mechanism design upgrades data governance from a one-time project to a systematic engineering process of continuous iteration, providing an actionable pathway framework for the large-scale implementation of industry models.

The seven key tasks are progressively deployed across four levels: "data—model—scenario—ecosystem." At the data and model level, the notice requires implementing entities to inventory data resources within their industries, and through data annotation and knowledge engineering, refine and form high-quality general knowledge datasets for the industry, with no fewer than five sorted out per industry. Based on these general knowledge datasets, at least one industry model mastering the industry's technical mechanisms should be developed, along with no fewer than five application cases. At the scenario and agent level, no fewer than 30 high-value application scenarios should be condensed per industry. For each scenario, at least one industry-specific knowledge dataset should be constructed, and a specialized model or characteristic agent should be built, with each agent requiring implementation in no fewer than three practical cases. China's comprehensive industrial system and complete industrial chain provide a natural scenario-rich mine for the large-scale construction of high-quality datasets—from real-time sensor data in steel blast furnaces to visual inspection data on automotive production lines, from operational monitoring data of power equipment to defect classification data of electronic components, each sub-sector accumulates unique tacit knowledge and professional mechanisms. The current industry pain point is that this data, scattered across different enterprises and production lines, has yet to form a reusable asset pool, and the construction of general knowledge datasets is the key lever to break through this barrier.

Regarding infrastructure and ecosystem support, the notice proposes for the first time the creation of "Model-Data Resonance" spaces, requiring each provincial-level region to build no fewer than three, and each central enterprise to build no fewer than one. Such spaces must possess capabilities for trusted cross-entity data integration, collaborative model training, and secure and compliant application, and are encouraged to interconnect with national data infrastructure, gradually evolving into "agent factories." The notice also calls for the formation of "Model-Data Resonance" innovation consortia, with at least one established per key industry, uniting entities from computing power, models, data, and applications to jointly conduct full-stack solution research and development. The initiative also establishes a "Key City" mechanism, where cities with a strong foundation in the AI industry can apply to become benchmark pilots. In terms of talent support, the notice proposes systematically cultivating multi-disciplinary composite talents proficient in industry applications, data science, and model mechanisms through organizing "Deep Dive" activities, building training bases, and implementing mechanisms like "Open Competition for Selecting the Best Candidates."

Regarding the timeline, the notice specifies key milestones from initiation to conclusion, with a tight schedule. By May 30, 2026, provincial-level industry and information technology authorities, in conjunction with data management departments, must compile and submit implementation plans. By August 30, a phased summary report must be submitted, and the Ministry of Industry and Information Technology and the National Data Administration will organize experts to conduct a mid-term evaluation. By November 30, an initiative summary report must be submitted, after which the two departments will evaluate the overall implementation and release the list of completed initiatives and cities. The Ministry of Industry and Information Technology and the National Data Administration will establish a unified platform to showcase various achievements and provide preferential support in relevant policies and projects for regions and enterprises with good implementation results.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com