Did you know? Semiconductor processes currently have a high share of defective output (about 10%)????? A semiconductor goes through 8 steps to be completed, and most of the defective ones are sorted out by people directly from images at the finished-product stage. It takes a lot of time, and it is hard to identify the cause of defects. That is why AIBIZ built this. Before they become finished products, AI predicts defects with deep learning and also identifies the causes of defects. After signing a supply contract with Company A, a non-memory semiconductor company, and carrying out the solution validation stage, there were outstanding results! - Defect rate 5.2% -> 0.1% - Defect analysis time 7 days -> 2 hours - Time to defect recognition 38 days -> 1 hour