Generative Adversarial Network

[ArmanZarei]

Generative Adversarial Network (First Phase)

[sarakh1999]

Hi, Thanks for preparing the notebook.
I'm not sure that I have seen all the content you prepared because some parts of the notebook were not shown properly.
I think you can ignore the mathematics and just tell the mechanism of GANs and how they work and the architecture simply. For example, you can add an example of a detective and a forger to produce money and explain the mechanism (how detector and forger make each other stronger:)) ) based on this example to seem easy for students (This is just a suggestion!).
You can also add a subsection for different types of GANs such as conditional GANs ( I guess you have already add it to your notebook, but it was not shown properly for me and I could not see accurately:(( )
You can add examples of these networks in different fields for example in computer vision. You can also write briefly about convolution and working with images and give an example. You can also write about attacks or the usages in NLP.
Please add a piece of code of generative models that produce new samples. If it is not challenging for you, it can be amazing to implement it on a dataset of faces to produce new fake faces. I guess students will find it amazing! MNIST dataset is somehow fine.
You can also search to find whether some famous companies such as Google, Microsoft, Facebook, ... use this network for different tasks in their AI section or not. If you find some interesting works implemented with GANs, you can add them to the notebook.
You can also write about most important articles using GANs and give their links.