Deep Multistage Multi-Task Learning for Quality Prediction of Multistage Manufacturing Systems
This work addresses quality prediction in manufacturing systems, which is an incremental improvement over classic modeling techniques that fail to account for correlations within or between stages.
The paper tackles the problem of predicting multiple quality indices in multistage manufacturing systems by proposing a deep multistage multi-task learning framework that jointly predicts all output sensing variables, resulting in superior performance compared to benchmark methods and offering interpretability through variable selection techniques.
In multistage manufacturing systems, modeling multiple quality indices based on the process sensing variables is important. However, the classic modeling technique predicts each quality variable one at a time, which fails to consider the correlation within or between stages. We propose a deep multistage multi-task learning framework to jointly predict all output sensing variables in a unified end-to-end learning framework according to the sequential system architecture in the MMS. Our numerical studies and real case study have shown that the new model has a superior performance compared to many benchmark methods as well as great interpretability through developed variable selection techniques.