Recently, Luculent was invited to attend the "Shangxin Zhizhen" Scientific and Technological Innovation Technology Forum hosted by China Mobile Shanghai Industrial Research Institute (hereinafter referred to as "Shanghai Research Institute"), and delivered a keynote report titled *AI+Manufacturing: Application and Practice Sharing of High-Value Scenarios*. Focusing on the theme of "AI+Manufacturing Integration and Development", the forum gathered industry experts and leading figures to jointly analyze the current development status of the industry, study the pain points and challenges in implementation, and look forward to future development trends.

The leadership team of Shanghai Research Institute, including Chairman Huang Gang, Deputy General Manager Liu Qinghua, Secretary Wen Xue, Deputy General Manager Qiao Jianshe, Deputy General Manager Yan Maosheng; Chief Technical Officer Luo Chenchi; Deputy General Manager of Industrial Internet Innovation Department Li Haiwei, General Manager of Platform Department Zhou Wei, Deputy General Manager of Application Department Yu Jianjun, as well as backbone personnel in the AI technology field, attended the meeting.

Cloud-Edge-End Collaboration: The Optimal Path for Industrial AI Implementation

Experts from Luculent pointed out that the current integration and application of AI in the manufacturing industry is not lacking in advanced algorithms and technologies, but faces in-depth problems in data, computing power, scenarios, collaboration and ecology. To truly realize the closed-loop implementation and large-scale promotion of industrial AI, the key lies in systematically breaking the five traditional technical bottlenecks of data, model, edge-end, control and trusted security.

In the future, enterprises, especially large group-level enterprises, will increasingly emphasize intensive management, and the "cloud-edge-end collaboration" management architecture will become the best practice position for industrial AI applications.

The "cloud" represents the group, serving as an empowerment center, scheduling center and supervision center. It gathers various expert experiences of the entire group to form a complete knowledge base, case base and model base, and distributes mature models to factories to empower frontline production. The "edge" refers to the factory side, which feeds back the models established, trained and optimized in production practice to the group, further enriching the group's knowledge system. The "end" includes various equipment, sensors and operation terminals on the frontline of production, which are responsible for collecting real-time production data and executing instructions issued by the models.

The "cloud-edge-end collaboration" model forms a complete large closed loop, enabling the group-level models to continuously self-learn and iteratively optimize in production practice. Experts said that if enterprise resources and data in the industry are integrated to build an industry-level "AI Industrial Brain", it will promote a leap in the intelligent level of the entire manufacturing industry.

Joint Efforts of "Large Model + Small Model" to Maximize Value

In the industrial field, it is necessary to build a collaborative system of "large and small models" that truly understands industrial scenarios and meets industrial needs, realize sustainable evolution, and enable AI to continuously optimize model capabilities through learning, transforming real-time data insights on the frontline into scientific business decisions for enterprises.

Experts vividly compared large models to "liberal arts students", which have strong understanding and decision-making capabilities, integrating and analyzing multi-dimensional information and issuing optimization solutions; small models are like "science students" with professional skills, focusing on perception and execution in industrial scenarios, accurately capturing production data and completing specific operation instructions.

Taking equipment fault diagnosis as an example, the small model for equipment fault diagnosis can monitor the operating status of equipment in real time and quickly detect potential faults; the large model, combined with information such as equipment maintenance and operation and maintenance data, enterprise production plans, and historical supply records of spare parts suppliers, automatically issues maintenance solutions that minimize the impact on production and reduce maintenance costs, providing scientific references for the operation and maintenance department.

Industrial AI + High-Value Scenarios: Significant Improvement in Quality, Cost Reduction and Efficiency Enhancement

The true value of AI lies in helping every industry re-improve efficiency, reduce costs and create new business models.

With the increasing popularity of unmanned factories, the production safety of factories will be greatly improved, and operational efficiency will also be continuously optimized. All these are profound changes in enterprise production and operation brought about by AI. Luculent suggests that industrial enterprises take the lead in applying AI in high-value business scenarios:

For example, in production safety, build safety lines for equipment, personnel and environment by using AI visual recognition, intelligent patrol inspection and other technologies; in production operation, ensure stable and efficient production processes through real-time monitoring and abnormal early warning of intelligent monitoring; in intelligent operation and maintenance, empower process optimization and quality optimization with AI, reduce management, manufacturing and maintenance costs, and improve production efficiency and cost control capabilities; in energy consumption optimization, optimize energy scheduling and energy consumption management based on AI technology to help enterprises achieve green economic development; in operation, connect the entire link of production, management and decision-making to realize efficient operation of business processes.

The core competitiveness of industrial AI in the future lies in the ability to understand industrial scenarios, the ability to precipitate industrial mechanisms, and the implementation ability to combine AI with systematic engineering. Only a system rooted in the frontline of production, understanding processes, and collaborating with large and small models can truly turn AI into new productive forces in factories.

AI Empowers the Entire Industrial Chain to Create New Productive Forces

AI is not just a technological change, but a competition of concepts, global capabilities and behaviors among enterprises.

Currently, Luculent is consolidating the intelligent foundation with AI native platform technology, reshaping scenario applications and reconstructing the value of the entire industrial chain with AI, creating a symbiotic and win-win industrial ecosystem through model innovation, and fully opening a new chapter of All in AI. Internally, it builds a global AI empowerment system to promote high-quality enterprise growth; externally, it creates a one-stop digital and intelligent platform to help customers build independent innovation capabilities.

Shanghai is actively promoting the "Modeling Shanghai" strategy for artificial intelligence. Experts from Luculent suggested that Shanghai Research Institute focus on AI+modern manufacturing and intelligent manufacturing, deepen the integration of AI and industrial scenarios, build a digital and intelligent team with in-depth interdisciplinary integration, make good use of large models and intelligent agents, strengthen the construction of high-quality industry data sets, build a capability system, combat system and ecological system, and achieve new breakthroughs in the field of "AI+New Industrialization".