Geely Changxing Automatic Transmission Co., Ltd. with machine arms waving at high speed and technicians busy working on the automatic assembly line. Photographed by Tan Yunfeng (WI Visual)

Artificial intelligence is a strategic technology leading a new round of scientific and technological revolution and industrial change, and has a pioneer effect of spillover. Data from the Ministry of Industry and Information Technology shows that there are more than 4,500 enterprises in the field of artificial intelligence in China, and innovations such as smart chips and general large models are accelerating, while digital workshops and smart factories are speeding up construction.

‘To give full play to the advantages of China's complete industrial system, large-scale industry, rich application scenes, with the deep integration of artificial intelligence and manufacturing as the main line, with intelligent manufacturing as the main direction of attack, and scene application as the traction, to promote the intelligent transformation of the manufacturing industry, and the high level of empowerment of the industrial manufacturing system.’ Tao Qing, director of the Operation Monitoring and Coordination Bureau of the Ministry of Industry and Information Technology, said.

Richer application scenarios

This year's ‘Government Work Report’ proposed to deepen the research and development of big data, artificial intelligence and other applications, to carry out ‘artificial intelligence +’ action. The implementation of manufacturing digital transformation action.

‘The manufacturing industry is the main battlefield of the “artificial intelligence +” action, China's manufacturing industry is huge, and the information infrastructure is perfect, laying a good foundation for the deep integration of artificial intelligence technology and manufacturing.’ Shan Zhiguang, director of the Department of Informatisation and Industrial Development of the National Information Centre, analyzed that in terms of digital infrastructure, the scale of China's arithmetic industry is growing rapidly, and the total scale of arithmetic has reached 230EFLOPS, ranking the second in the world, and the key core technologies such as 5G and 6G are constantly making breakthroughs, and high-performance computing continues to be in the world's first echelon; in terms of the industrial foundation, China has the most complete industrial system in the world. The overall scale of the manufacturing industry has remained the first in the world for 14 consecutive years, and the scale of the core AI industry has been growing.

Wu Tongning, deputy director of the Artificial Intelligence Research Institute of the China Academy of Information and Communication Research, said that China has the supply advantage of a complete industrial system, and the integration of AI and manufacturing can further enhance the competitiveness of the industrial chain supply chain by optimising the level of supply and improving the performance and quality of products. At the same time, the vast field of manufacturing provides a rich application scene for the development of artificial intelligence.

It is also important to see that AI technology faces many challenges in the process of integration with the manufacturing industry. One hand, there is still a gap between the chip R&D design and process and the international advanced level, the artificial intelligence standardisation system needs to be established urgently, and there is a shortage of core high-end professionals; on the other hand, the data security problem is also becoming more and more prominent, and how to protect the core data and intellectual property rights of enterprises has become a topic of concern for all parties.

‘The next step is to focus on algorithms, arithmetic and other large model underlying technology, to accelerate the promotion of intelligent chips, large model algorithms, frameworks and other basic key core technology and product breakthroughs; to establish and improve the standard system of AI-enabled new industrialisation; and to improve the mechanism of AI talent training, safety and security.’ Tao Qing said.

In Wu Tongning's view, it is crucial to promote standardisation. AI-related industry standards should be established and promoted through the establishment of an AI standardisation technical committee, technical research and validation should be carried out to promote the unity and interoperability of technology, and the deep integration of AI products and applications with industry should be strengthened.

In terms of talent cultivation, Shan Zhiguang suggests promoting the construction of a multi-level talent cultivation system covering high-level talents, professional and technical talents, industry skilled talents, and primary and secondary school AI basic education talents; strengthening new types of skills training, and guiding laborers to enter AI industry-related positions such as data labelling; and actively carrying out international exchanges and cooperation in digital education, and striving to cultivate global governance and internationalisation adapted to the development of the times innovative talents to adapt to the development of the times.

More obvious empowering effect

The manufacturing industry is the pillar of the national economy and has a certain digital foundation. AI technology can shorten the R&D cycle by optimising the R&D and design process; improve production efficiency and product quality control by enhancing the intelligence of the production line; and reduce the equipment failure rate by carrying out fault diagnosis and preventive maintenance based on historical data and expert files.

‘Artificial intelligence technology can accurately predict market demand and the iterative trend of technology products, which in turn can assist enterprises in formulating long-term development strategies, promote the research and development and application of new technologies, and support continuous innovation.’ Wu Tongning introduced that the intelligent logistics management system can achieve optimised logistics routes and vehicle scheduling to ensure that goods can be delivered in the fastest and most cost-effective way. In addition, the use of artificial intelligence technology to carry out predictive maintenance of equipment, combined with industrial Internet and 5G technology, real-time monitoring, access to equipment operation data, projecting the time of the emergence of equipment failures, reducing the risk of production interruptions, and bringing higher stability and reliability to the manufacturing industry.

Tao Qing said that the Ministry of Industry and Information Technology will deepen the integration and application of AI technology in the whole process of manufacturing, and significantly improve the level of R&D, pilot, production, service, management and other aspects of intelligence. Aiming at key industries with great influence on the national economy, strong driving ability and good digitalisation foundation, it will carry out special actions for AI-enabled new industrialisation.

‘To achieve intelligent upgrading of key links, key industries and key products in the manufacturing industry, it is necessary to do a good job of scene mining and business process standardisation.’ According to Shan Zhiguang, key links such as production, quality inspection, inventory management and supply chain optimisation are often characterised by high repetitiveness, large data volumes and complex decision-making, which are suitable for the application of AI technology. For example, the use of machine vision technology for automatic inspection in the quality inspection process.

Each key process in the production of display panels requires AOI (Automatic Optical Inspection) equipment to take pictures and identify relevant defects. Previously, defects were classified manually, and the entire process, with more than 100 processes, was labour-intensive and time-consuming. In order to reduce labour costs, Gertrudong Zhi developed an AI visual inspection system in conjunction with TCL Huaxing, which is based on artificial intelligence technology to identify and classify pictures. While significantly improving the inspection efficiency, it reduces the number of inspectors by 90% and can inspect more than 3 million pictures a day. In addition, AI technology can effectively avoid problems such as easy fatigue of personnel and cognitive differences between people, and further improve detection accuracy.

Better industrial ecology

In the process of manufacturing upgrading, the power of a single technology is limited. The application of artificial intelligence in enterprises needs to be combined with 5G, cloud computing and other technologies to maximise its effectiveness.

"Firstly, a high-speed 5G network, large-scale cloud computing resources and an advanced AI base need to be established. Secondly, a unified data management platform can integrate data from production lines, supply chains, markets and other areas, which can be stored, processed and analysed via cloud computing to provide the necessary input for AI algorithms. High-speed network infrastructure can be used to build a data transmission bridge between manufacturing equipment and the data centre, ensuring real-time data transmission and providing a solid guarantee for equipment feedback of large model training and inference results." Wu Tongning said.

An Xiaopeng, vice president of AliCloud Intelligence Group, also believes that the ‘public cloud + AI’ as the representative of the technology system all-round empowerment and support is another obvious feature of the current manufacturing transformation. Public cloud is the optimal path to break through the bottleneck of high-end chips, through efficient connection of heterogeneous computing resources, break through the bottleneck of a single performance chip, and collaborate to complete large-scale intelligent computing tasks. To ‘public cloud priority’ strategy as an important element of the digital transformation of the manufacturing industry related policies and planning, clear medium- and long-term development goals, key tasks and safeguard measures; to avoid the chip ‘squeeze’ phenomenon, vigilance around the ‘small scattered more’ in a rush. Small scattered more than’ a rush to build the arithmetic centre; to data centre utilization efficiency as a data centre construction assessment indicators, reversing the data centre “heavy construction, light operation” “heavy investment, light performance” construction mode.

Artificial intelligence empowered manufacturing industry also needs a good industrial ecology as support. ‘Attention should be paid to the important role played by industrial alliance organisations in promoting industrial development, facilitating research and innovation, strengthening cooperation and exchanges, participating in policy and standard setting and improving international competitiveness.’ Wu Tongning said.

According to reports, the China Artificial Intelligence Industry Development Alliance (hereinafter referred to as the ‘AIIA Alliance’), jointly initiated by the China Academy of Information and Communication Research and industrial units, has more than 1,000 members, and has made important achievements in building a platform for cooperation between industry, academia, research and application, constructing China's AI industrial ecosystem, and strengthening the in-depth fusion of AI with various economic and social fields. Since 2023, the AIIA Alliance has launched the collection of ten pioneering cases every year to select typical application practices with high value, benchmarking, and real realisation in the industry, so as to better promote the application of AI technology and the development of AI industry in China.

Artificial intelligence open source ecological construction has always been the industry's key direction of attention. ‘It is necessary to explore the construction of an open source ecosystem with Chinese characteristics, encourage technology companies to join the construction of open source ecosystems, promote open source communities, open source foundations and other platforms to collaborate with colleges and universities, enterprises, research institutes and other institutions to conduct research; open source competitions are held to select outstanding talents, showcase innovative achievements, and spread the popularity of open source concepts.’ Shan Zhiguang said.

Source: Economic Daily