Fast model-driven target optimization methods using machine learnings

The goal of this paper is to explore machine learning solutions to improve the run time of model-based retargeting in the mask synthesis flow. The purpose of retargeting is to re-size non-lithography friendly designs so that the design geometries are shifted to a more lithography-robust design space. However, current model-based approaches can take significant runtime. As a result, this step is rarely done in production settings. We propose a machine learning solution that moves segments locally like optical proximity correction (OPC). In addition, we propose a convolution neural network (CNN) that takes in the target layer as an image and outputs the new target as an image. Finally, we discuss the experimental results where we show that we can achieve an order of magnitude runtime improvement while maintaining similar accuracy to traditional model-based retargeting techniques.