Smart Diagnostics for 3D CFET: A Machine Learning Approach to Failure Analysis

This work introduces a novel Convolutional Neural Network for classifying transfer characteristics in emerging Gate-All-Around MOSFET. Trained on vast experimental dataset, the algorithm successfully identifies distinct failure modes across wafers with complex processing variations. The automated analysis enables faster yield enhancement and process optimization for next-generation 3D MOSFET technologies.