CSE 6240 Project #1
Blog Series on Generative Adversarial Networks (GANs)
Generative Adversarial Networks are an unsupervised learning technique capable of generating
realistic synthetic data, like images. For this assignment you are to write a two-part blog tutorial
on GANs. For reference on blog style you may want to refer to:
http://colah.github.io/posts/2015-08-Understanding-LSTMs/
http://karpathy.github.io/2015/05/21/rnn-effectiveness/
And for an introduction to GANs you can look at:
https://arxiv.org/abs/1701.00160
http://blog.aylien.com/introduction-generative-adversarial-networks-code-tensorflow/
Part 1 -- Background and How They Work [50 points]
This part should introduce the reader to the history and context (applications) of GANs, and
explain how they work. Grading:
● 20% Style
● 30% Background
● 50% Explanation (clarity, completeness)
Part 2 -- Application [50 points]
For this part you should create a toy or real application using GANs, and explain the problem,
your approach, results and future work / conclusions. Grading:
● 25% Originality & appropriateness of the problem
● 50% Solution correctness, completeness and explanation
● 25% Discussion and conclusions
Deliverables
Your blog posts + supporting files and data + your code as python | ipython notebooks.
