Polar mesospheric clouds (PMCs) have been studied for thirteen years by NASA's Aeronomy of Ice in the Mesosphere (AIM) satellite. The Cloud Imaging and Particle Size (CIPS) instrument onboard AIM has taken many images of PMCs over this time. Such a large number of images makes CIPS data ideal for training neural networks which require large datasets. CIPS images were used to train a Generative Adversarial Network (GAN) to train towards being able to generate purely artificial CIPS-like images.
Hedges, Austin, "Making Artificial CIPS Data With a Generative Adversarial Neural Network" (2020). Fall Showcase for Research and Creative Inquiry. 77.