In order to predict the reliability of high electron mobility transistors (HEMT) devices under high-power microwave (HPM) stress, a deep learning algorithm network model was constructed to predict the device lifetime. The influence of HPM stress on HEMT was studied using computer-aided design (TCAD) technology. The results show that the relative error percentage of the prediction results of the deep learning algorithm is less than 15%, and the relative error percentage of most predicted values is less than 5%. Comparative experiments with five traditional machine learning methods (support vector machine, decision tree, K-nearest neighbor algorithm, ridge regression, and linear regression) indicate that the deep learning algorithm has the best performance, with the minimum average error percentage. This data-based deep learning algorithm model not only enables researchers who are not familiar with semiconductor devices to quickly obtain the lifetime data of the devices under any conditions; but also can be used as a data-driven device model to reflect the HPM reliability of individual devices and applied in device design. The application of deep learning in the field of device lifetime prediction has an extremely broad prospect in the future.
Su et al. (Fri,) studied this question.