Hand in hand with corporation DataGrid development "" to judge product defects by AI technology
To strengthen manufacturing ~ by manufacturing field digital transformation
Our company with the IoT r&d center, established in 2017 as the core, to push the digital transformation of manufacturing site (DX), is committed to using IoT/AI technology to improve productivity and safety.One of the measures is to determine defects can be development of AI, the AI can replace skilled workers, automatic product appearance checks.But in order to make a high standard of defect judgement of AI study to determine the benchmark, must be prepared to provide a variety of modes of learning data.Collect learning data to spend a lot of time and cost, therefore, from the actual situation in the company, led by the manufacturing companies has yet to fully use the AI.
On the subject, in the real study data is very limited, the measures to develop can match skilled defect judgement ability of AI.
By "defect judgement technology", our company have predicted could be AI misjudgment difficult mode, through the corporation of the DataGrid has "suspected defect generation technology" to generate a difficult pattern of bad data.And repeated "difficult model to overcome the learning technology training cycle (weakness)", namely through the defect in the generated data to determine the AI again to study, to overcome the difficult mode, improve accuracy of defect judgement AI.
Joint development of the technology in the process, what place is worthy of attention?
We recognize that the faulty judgement and suspected defects generated AI AI synchronous growth of technology has not yet been established, so if you can develop this technology, it is a global initiative.
Tease out after students attend mock exam is not good at subjects, key around are not good at subjects were specially trained, can efficiently improve the score.Likewise, through this technology makes the computer automatically repeated weaknesses training cycle, it is possible to let the development speed of defect judgement AI improved dozens of times, or even hundreds of times.In addition, in addition to the image data, in this difficult model to overcome the learning technology can also be applied to various headed by data of time sequence data.In addition to the appearance inspection process automation, this technology can be applied to predict equipment failure and inventory management and so on all kinds of the scene of the manufacturing process, which would accelerate the process of manufacturing field of automation.We will use these technologies, is committed to enhance manufacturing capabilities.
