Giters

vt-vl-lab / paper-gestalt

Deep Paper Gestalt

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

The use of GAN

Zengjie12 opened this issue · comments

commented

Dear author:
I am a new learner for generative adversarial networks.I am very interested in your research.Here,I want to ask you for some question about GAN：
1.I do not clearly know what the GAN used to do?I guess that it is used to generated what the good paper.
2.What the methods you have used to classify the paper which is good or not?
3.Do you consider that GAN can be used to the simple binary classification problem,which means just learn datas and predict it belongs to 0 or 1?
I am sorry to bother you to ask the primary question.I am very appreciated that you can answer my question.

commented
  1. You can check out the 2016 tutorial on GAN: https://www.youtube.com/watch?v=HGYYEUSm-0Q
  2. Just finetuning ResNet-18 on the Computer Vision Paper Gestalt dataset
  3. Yes, when you train a GAN you will also learn a discriminator that classify whether a paper is real or not. However, this discriminator does not aim to classifier whether a paper is good or bad.

Hope this helps.