There is absolutely no doubt that artificial intelligence is subverting the global manufacturing landscape. According to IDC estimates, Taiwanese manufacturing will import artificial intelligence by 25% in 2018. At first glance, this figure is not amazing, but in fact, manufacturing is the area where Taiwanese industry has introduced the most artificial intelligence technology after the financial industry. How should manufacturers follow and stay out of this trend? Before introducing artificial intelligence technology, manufacturers must understand that artificial intelligence relies on data foundation, through the feeding and digestion of data, and finally allows artificial intelligence to achieve self-adjustment and optimization, so the data can be said to be artificial intelligence. "Basic food." Lin Weiren, associate director of the Far East Advanced Fibers Operations Department, actually uses the application-side analysis. Some manufacturers have insufficient data. Therefore, even if artificial intelligence is introduced for analysis, the accuracy rate is less than 60%. If the original human level cannot be achieved, artificial intelligence will be Meaningless. Therefore, Lin Weiren suggested that manufacturers must have big data support before artificial intelligence. If operators have the idea of ​​actively introducing artificial intelligence, they must first prepare data. The more they accumulate, the more accurate the analysis will be. Liang Yuci, deputy general manager of the IBM manufacturing group in Taiwan, analyzed that the current domestic manufacturing industry's use of artificial intelligence mainly focuses on two needs, one is quality inspection, and the second is equipment predictive maintenance. At the current stage of the industrial type, the manufacturing plants are still dominated by manpower operations, as if they were manually responsible for the quality inspection process. However, this is often caused by the quality of the eyes caused by fatigue, or the naked eye can not pick out too small flaws. Even because of the sharp decline in the labor force, resulting in insufficient manpower or high costs, according to this trend, the inspection of artificial products will be difficult to respond to market demand. Therefore, the visual inspection technology that replaces human resources with machines will play a very important role. Among them, artificial intelligence technology is a tool that enables the machine to evolve rapidly and improve the quality and efficiency of testing. The difference between artificial intelligence and traditional AOI is that the former does not need to write algorithms in a timely manner. As long as the system discriminates through a large number of images, what is good, what is bad, and what are their characteristics, artificial intelligence will encourage the machine to continuously learn and Optimization, in addition to being able to achieve faster detection efficiency than people, the accuracy will be more accurate, and finally reach a level that is incomparable to the human eye. Thanks to the maturity of deep learning, image recognition technology has been able to perform quite well in recent years. IBM pointed out that the current accuracy of using artificial intelligence for quality inspection can reach more than 90%, and it helps to reduce the detection time by 80% for auto manufacturers. Visual inspection can be said to be the most extensive field of artificial intelligence technology used in the manufacturing industry, which not only greatly increases the yield, but also reduces the manpower demand for online production. In terms of long-term performance, once the benefits are reduced by manpower reduction, manufacturers can clearly feel the significant difference between smart manufacturing and traditional models. To be fair, this is the most direct and fastest way for the industry to cost down. Liang Yici observed that visual inspection should be the first application that most manufacturers will take the lead in introducing artificial intelligence. As for the pre-knowledge maintenance of equipment, although its structure is huge, it is necessary to integrate data analysis of non-structural data such as sensors and parts. However, for the semiconductor or high-tech electronics industry, the loss caused by the machine half-stop is not allowed. Therefore, the industry has already invested in research and development long before the equipment foreseeing prevention issues have yet to be discussed, and the development has now had a relatively superior technical level. In order to strengthen the upgrading of manufacturing technology, artificial intelligence has become a new manifestation. This is a trend that the industry will avoid in the future, but there are still many differences arising from the different types of industries. For example, in the case of equipment for prevention and maintenance, this application emphasizes the ability to achieve real-time monitoring and early warning for machine equipment. For the high-tech electronics industry that produces online precision electronic components, it can ensure stable operation of the production line. As electronic parts become more sophisticated, the requirements of the high-tech electronics industry for quality inspections are increasing. But this does not mean that the traditional industry does not care about quality control, but it is more eager for the adjustment of manpower demand than the manufacturing mode of traditional industries. Liang Yici added that the problems that the traditional industry is more likely to face in the future are a small number of challenges. She pointed out that if the demand for production changes rapidly, the traditional industry will change the production line content more frequently to meet the diversified needs, which will cause the machine to have a larger amount of data and more diverse when introducing artificial intelligence. The learning model can adapt to the ever-changing market demand. The issue of talent change is another challenge for traditional industries. Most of the factory masters in the traditional industries have done more than a decade. Because the old habits are difficult to change, many professional knowledge and valuable experience are stored in the mind, not systematic preservation. Coupled with the sharp decline in talent, traditional industries have to introduce new technological elements to solve this problem. IBM's IoT Equipment Advisor, similar to IBM's Watson, uses a natural language question answering system to confirm the credibility of answers based on acquired knowledge. Equipment Advisor will use the accumulated data to judge the user's problems, recommend the most reasonable method, and give confidence index to help users to assist in the repair, maintenance, procedures or technology related problems. decision making. To put it simply, it is to specify and digitize the precious experience of the factory's masters for 20 or 30 years, and to provide assistance through the presentation of any mobile device. This can be regarded as the category of knowledge management. Liang Yici believes that this is also an artificial intelligence application that the traditional industry has the opportunity to contact and actually operate in the future. Since this has less relevance to automation, even if the industry does not have automation technology at this stage, or whether the device has network connectivity, there is no conflict between the two. For the traditional industry, it is even faster. Imported artificial intelligence application. Battery Cable,Lithium Battery Cable,Battery Cable Connector,Silicone Cover Wire Harness Dongguan ZhiChuangXing Electronics Co., LTD , https://www.zcxelectronics.com