An application of neural networks to prediction of SL laser packagecoupling stability characteristics

摘要

Artificial intelligence (AI) is fast becoming a practical and costeffective approach to data and sensor based decision making processes.This report discusses the application of neural networks, a branch ofAI, to device quality evaluation. The specific product considered is theSL-560 laser package. We demonstrate the application of neural networksfor making decisions on device quality, based on observation ofmanufacturing process data patterns. The current work aims at usingneural networks to model the opinion of human experts. These experts arepart of a pedigree review group which examines a large variety of dataon each device and then assigns a grade to each device. The specificcharacteristic considered here is the coupling stability of the laserpackage through manufacturing process