Methods for training machine learning model for computation lithography

A computer program product comprising a non-transitory computer-readable medium having instructions therein, the instructions, upon execution by a computer system, configured to cause the computer system to at least:

• obtain (i) a trained process machine learning model configured to predict an integrated circuit pattern on a substrate as produced using a semiconductor patterning process, and (ii) a target pattern;
• train a mask pattern machine learning model configured to predict a mask pattern that is configured to interact with radiation to create an image from the radiation to expose a resist layer on the substrate using the semiconductor patterning process,
• wherein the training includes:
• predicting, by the mask pattern machine learning model, the mask pattern from the target pattern,
• processing the mask pattern using the trained process machine learning model to generate a predicted integrated circuit pattern,
• evaluating a cost function that determines a difference between the predicted integrated circuit pattern and the target pattern,
• updating parameters of the mask pattern machine learning model such that the cost function is reduced, and
• repeating the predicting, processing, evaluating and updating until the cost function is reduced relative to an earlier iteration, thereby reducing the difference between the predicted integrated circuit pattern and the target pattern,
• wherein the mask pattern machine learning model provides optical proximity correction and/or source mask optimization of the mask pattern.