> You can add information, by making rules like the class doesn't change if the thing is a different color or size or position, or on a different background. Bu

> You can add information, by making rules like the class doesn't change if the thing is a different color or size or position, or on a different background. But you can't automatically create new images from a distribution that has been learned and expect them to add information. I am not quite sure I follow. Why would any of those things be true for all models?