Fréchet inception distance measures the quality of this type of data given two distributions of this data. The Nightshade tool adds noise to this kind of data to “poison” models that train on it. A model designed for this type of data introduced skip connections to enable deeper models. A 2018 study by Buolamwini and Gebru found evidence of discriminatory disparities in classifiers for this type of data. In an architecture primarily (10[1])used for this type of data, pooling steps reduce dimensionality between operations that move a filter by a specified stride. A collection (10[1])of this type of data developed by Fei-Fei Li was used for the ILSVRC competition. AlexNet and ResNet use convolutional neural networks to classify this type of data. For 10 points, name this type of data generated by models such as DALL-E and Midjourney. (10[1])■END■