.png){kind=icon}
.png){kind=icon}
Return
Data-driven optimization of accessory combinations for final testing processes in semiconductor manufacturing
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jmsy.2022.03.014.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper proposes an integrated framework to predict first yield, in order to find the best-practice accessory combination for the final testing (FT) process in semiconductor manufacturing. Firstly, the entity embedding method is adopted to convert the categorical data of accessory combination into multidimensional vectors. Several possible algorithms in machine learning are then examined to establish the yield prediction model and select the one that is the best fit. Once the best machine learning model has been determined, the genetic algorithm (GA) embedded in the yield prediction model is used to search for the best-practice accessory combination. This is the combination that gives the highest first yield estimate. The synergy between machine learning and the search heuristic produces an intelligent predictive system that can circumvent the adverse yield rates generated by using inappropriate accessory combinations. The Overall Equipment Effectiveness (OEE) of the FT process can be maintained in a highly stable condition.
Keywords:
Semiconductor manufacturing
Machine learning
Accessory combination
Entity embedding
Final testing
Journal
IF:
14.2
Papers:
2.7K
Citations:
1.6W