Test pattern generation systems and methods

A test pattern generation system, comprising:

• before training a generator of a test generation machine learning circuitry, receiving, by rasterization circuitry, a plurality of random noise vectors;
• and
• before training the generator of the test generation machine learning circuitry, generating by rasterization circuitry, the plurality of random noise images as a plurality of raster images based on the plurality of random noise vectors;
• test generation machine learning circuitry including a generator and a discriminator, the generator configured to generate a pattern image based on a noise image, the discriminator configured to classify the pattern image as representative of a valid layout clip or an invalid layout clip, wherein training the generator of the test generation machine learning circuitry includes:
• providing the plurality of raster images as training input to the generator;
• generating, by the generator, first pattern images based on the plurality of raster images;
• generating, by the discriminator, loss information based on the first pattern images generated by the generator, the loss information indicating deviations between classifications of the first pattern images by the discriminator and a valid layout clip classification;
• and
• modifying parameters of the generator based on the loss information.

A method, comprising:

• before training a generator of a test generation machine learning circuitry, receiving, by rasterization circuitry, a plurality of random noise vectors;
• before training the generator of the test generation machine learning circuitry, generating by rasterization circuitry, a plurality of random noise images as a plurality of raster images based on the plurality of random noise vectors;
• providing a plurality of random noise images as training input to a generator of a test generation machine learning circuitry;
• generating, by the generator, first pattern images based on the plurality of random noise images;
• generating, by a trained discriminator, loss information based on the first pattern images generated by the generator, the loss information indicating deviations between classifications of the first pattern images by the discriminator and a valid layout clip classification;
• and
• modifying parameters of the generator based on the loss information.