DEFECTS DETECTION FOR A SEMICONDUCTOR COMPANY
Early failure detection, defects identification and fault diagnosis are among the most challenging problems in industrial management. The origin of process failures or defective components might be ambiguous and hard to investigate even by domain experts. Nowadays this activity can be effectively supported by data-based monitoring systems providing sensors data on the process status which can be analyzed in a multivariate fashion.
Business need
Detect the defective components at an early stage to reduce the cost of the testing process.
Challenge
Understand how defectiveness at the various testing phases can be explained through sensors measurements.
Solution
Develop an artificial intelligence approach to analyze if and how defective parts can be identified.
Benefits
Reduce the total cost of testing through timely defects detection. Better understanding of the production process by determining the root causes of abnormal components. Develop novel insights to drive manufacturing actions and improve the testing environment.