# What is the branch of data science that deals with generating data? Looking for tutorials

So far all the tutorials (Python/Tensorflow) I have been following dealt with processing the data, they are classification problems. Things that give Yes or No answers. Now I want to find tutorials on actually creating things using ML from large data sets.

For example lets say the task is to generate a face using millions of face photos. This is not a problem that produces statistics from data, but rather a problem of using statistics to generate data. My understanding is, the algorithm would generate random images and run them through the classifier until they closely resemble the face. Is this AI? I wouldn't consider it to be intelligent anything. I understand there is also a ton of creativity that goes into generating the images. But let's say I just wanted to use a +-RNG on every pixel. Is there an adaptive process where it would start matching at higher convolutional levels, where the image is just blobs of information and then use that information, adding or subtracting color, edges, etc, to move to lower levels that more closely resemble a face.

What is this type of problem called and what should I be looking for when searching a tutorial. If there is one that shows you how to do something similar, can you point me to this?

Thank you

$$p(Y|X)$$
Generative models are interested in $$p(X)$$ i.e. how is the data generated. Once the model learns this, it can produce "fake" images that look like your inputs.