PaperNova
Published PapersLearnPaper → CodePlagiarismPricing
LearnGANs (2014)
0%

14 lessons · 2014

Generative Adversarial Nets

1

The Generative Modelling Problem

15 min

2

From Noise to Data

15 min

3

The Discriminator

15 min

4

The Minimax Game

20 min

5

The Optimal Discriminator

20 min

6

The Global Optimum and Jensen-Shannon Divergence

25 min

7

The Training Algorithm

20 min

8

The Non-Saturating Generator Loss

15 min

9

Backpropagation Through Both Networks

25 min

10

Training Dynamics and Convergence

20 min

11

Failure Modes: Mode Collapse and Vanishing Gradients

20 min

12

Evaluating Generative Models

15 min

13

Results and the Latent Space

15 min

14

The Legacy: DCGAN, WGAN, StyleGAN and Beyond

15 min

Progress

0 / 14

Lesson 06 / 14

The Global Optimum and Jensen-Shannon Divergence

Theorem 1: C(G) = -log 4 + 2 · JSD(p_data || p_g)

Sign in to keep going

Create a free account or sign in to continue. Your progress stays saved on this device.

Create free accountSign in