Deep multistage multi-task learning for quality prediction of multistage manufacturing systems

In multistage manufacturing systems, modeling multiple quality indices based on the process sensing variables is important. However, classic technique predicts each variable one at a time, which fails to consider correlation within or between stages. We propose deep multi-task learning framework jointly predict all output in unified end-to-end according sequential system architecture MMS. Our numerical studies and real case study have shown that new model has superior performance compared many benchmark methods as well great interpretability through developed selection techniques.