Artificial intelligence is accelerating the "embrace" of the real economy
Artificial intelligence is accelerating deep integration with the real economy. Recently, the "2018 Artificial Intelligence and Real Economy Deep Integration Development Forum" has revealed many cases for us
The forum published the results of the “2018 Artificial Intelligence and Real Economy Deep Integration Innovation Project Application Work”, and 106 projects stood out in 2,428 applications. It is worth noting that the traditional industries in the application projects account for a considerable proportion
“In the application category, technology companies account for 10%, while traditional industries are up to 90%. Midea, Yili, Shanghai Aircraft Manufacturing, Yangtze Power, Sino-Japanese Friendship Hospital, etc., have formed exemplary cases in various segments. Zhao Ce, director of the High Technology Division of the Science and Technology Department of the Ministry of Industry and Information Technology
Artificial intelligence has shown obvious advantages for solving complex production problems in traditional industries and production problems under multiple constraints
Take the aircraft manufacturing industry as an example. 8 major systems, 28 key coordination interfaces, 33 key functions, 4.5 million parts... Wei Yingwei, Chairman of Shanghai Aircraft Manufacturing Co., Ltd. listed a series of figures on the forum to illustrate the complexity of aircraft design. The big plane is called the "industrial crown." Massive data, frequent human-computer interaction, and highly sophisticated algorithmic software systems have brought intelligent manufacturing scenarios to aircraft manufacturing
What is smart manufacturing? "Intelligent machines + advanced analysis tools + human-computer interaction is intelligent manufacturing." Chinese Academy of Engineering academician He Hezhen said, "Artificial intelligence, together with cloud computing, big data, 5G technology, has become an information technology that supports intelligent manufacturing."
According to Wei Yingxi, Shanghai Aircraft Manufacturing Company is building a 5G park and deploying a universal network covering the entire campus. “5G facilitates the interconnection of devices and real-time data collection, laying the foundation for big data for intelligent manufacturing.” At the same time, a group of advanced intelligence The use of equipment frees the hands of workers, such as automatic painting robots, flexible rail robots, automatic drilling and riveting robots, and laser scanning robots
On this basis, artificial intelligence plays a role in the use of machine vision, image recognition, voice interaction, big data analysis and other technologies to achieve smart perception, smart scheduling, smart detection, intelligent decision-making
Take the aircraft composite structure inspection as an example. “The traditional inspection process involves the manufacture of standard contrast sections for pre-embedded typical defects, scanning the structure of composite materials, forming images and models similar to X-ray films, and performing manual comparisons and evaluations.” Wei Yingxi said that due to the difficulty in making the plates, the cost is high. The aircraft structure is complex, the types of defects are numerous, the image signal is complex, and the rapid and reliable detection of aircraft composite structure has always been a worldwide problem.
The emergence of artificial intelligence has made things happen. “We collect massive data for samples of typical defects, build a database after extracting key information, and use 5G technology to store on the cloud platform and connect to the evaluation system. The established in-depth learning model can detect defects and iterate continuously.” Wei Yingwei According to the introduction, the current evaluation time has been shortened from 4 hours to a few minutes, and the cost of professionals has been reduced by 95%.
In fact, the use of artificial intelligence technology for smart detection, resulting in improved yield, can bring considerable profits to the enterprise. “Every 1% increase in the yield will increase the profit of several hundred million.” Zhang Wei, secretary of the board of Shenzhen Huaxing Optoelectronic Technology Co., Ltd. said. As a panel company, Huaxing Optoelectronics uses artificial intelligence to quickly learn and train large images of panels, and establishes high-precision models to achieve independent quality inspection. This technology, combined with the automated production line, from the input of materials to the finished product, does not require human intervention for a full two weeks, "production efficiency can be increased by 5%."
Hehe listed us more points of integration between artificial intelligence and the real economy. "Traditional visual equipment misjudgment rate may be as high as 20%, people will make repeated mistakes, but artificial intelligence can reduce the false positive rate to below 3% by learning the mistakes."
“Next, we will talk about the application of artificial intelligence technology in optimizing production parameters. The rubber source, processing plant, batch number of thousands of complex factors will affect the quality of rubber. After the introduction of artificial intelligence, the average pass rate of rubber compound will increase by 3%. -5%, an annual increase of 10 million yuan in profits."
Although the integration of the real economy and artificial intelligence has had initial results, there are still a lot of problems to be explored. A large amount of data, rich application scenarios, sufficient computing power, and active users are essential for the use of artificial intelligence, but these conditions are still far behind for many enterprises.
"We have visited Jiangsu and Zhejiang and found that many traditional real economy enterprises are still in the very early stage of digitalization. The data acquisition capability is still relatively weak, and the sensors are relatively general," said Xu Weiyuan, a partner of Paradise Silicon Valley.
Experts also recommend not to blindly pursue big data. "No matter how much data there is, there is still a lot of work to be done from the value of conversion. The data needs to be coordinated with a large number of projects from collection to entry. Enterprises must be accounted for and have cost constraints." Feng Junlan, chief scientist of China Mobile Communications Research Institute, said that enterprises It should also be considered that the application of artificial intelligence technology will have a great impact on the original operating system.