Prognostics and Health Management of Wafer Chemical-Mechanical Polishing System using Autoencoder

The Prognostics and Health Management Data Challenge (PHM) 2016 tracks the health state of components of a semiconductor wafer polishing process. The ultimate goal is to develop an ability to predict the wafer surface wear and tool settings through monitoring the components as the tool degrades overtime. This translates to cost saving in large scale production. The PHM dataset contains many time series measurements being under utilized by traditional physics based modelling approach. On the other hand, applying a data driven approach such as deep learning to this dataset is non-trivial. Unavailability of class labels is a main drawback to apply supervised deep learning methods, also for the application of unsupervised deep learning methods the feature space is not specifically targeted at the predictive ability or regression. In this work, we propose class labeling using the autoencoder based clustering whereby the feature space trained is found to be more suitable for performing regression. This is due to having a more compact distribution of samples respective to their nearest cluster means. We justify our claims by comparing the performance of our proposed method on the PHM dataset with several baselines such as the autoencoder as well as other state-of-the-art approaches.