Active Learning for Overlay Prediction in Semi-conductor Manufacturing

Abstract

In the manufacturing of semi-conductor devices there is a constant demand for increasing precision and yield. Measuring and controlling overlay errors is essential in this process, but these measurements are difficult and costly. Predictive models can be used as an addition to measurements, but they required labelled data for training. To achieve maximal performance with few measurements, active learning methods are explored that apply a sampling strategy to select which wafers to measure. The predictive model is a partial least squares regression, which is also used to provide informative visualizations of the high-dimensional data.