GAN C1W4: Can we increase Z space dimension so as to solve entanglement

Course Q&A Generative Adversarial Networks (GANS) Build Basic Generative Adversarial Networks
1

Since Z space doesn’t have enough dimensions relative to the number of features, can we increase Z space dimension to solve the entanglement?
Thanks for any response.

2

Hi,

I have moved this post to the GANS course 1 Week 4 category as other learners taking the same course might benefit from this.

Make sure if you have any course-specific doubts, explore the specialization category and post in the relevant course subcategory as course-specific mentors are actively answering the queries there and the threads in the general discussion might not come on their radar and remain unanswered at times.

3

Hello garynlfd,
z space entanglement can be defined as the criteria where the modifications of one feature to produce new image, then one or two other features also get altered. This occurs when the features in z space are uncorrelated. z space dimesion altering may not helps us to solve the issue in few cases. It is always prefered to first disentangle the features.

1 Like