I have a quick question regarding the use of different latent spaces to represent a distribution. Why is it that a Gaussian is usually used to represent the latent space of the generative model rather than say a hypercube? Is it because a Gaussian has most of its distribution centred around the origin rather than a uniform distribution which uniformly places points in a bounded region?

I've tried modelling different distributions using a generative model with both a Gaussian and Uniform distribution in the latent space and the Uniform is always slightly restrictive when compared with a Gaussian. Is there a mathematical reason behind this?

Thanks in advance!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.