I know that in general, a function $g$ can be a good auto-encoder (i.e., $g(x) \approx x$ for $x \sim D$) and on the same time $I(g(x);x)$ is small. This is the case when $g$ forms a good correlation but no compression of the data.

I'm curious if there is any work that suggests that the training process or some regularization tends to implicitly select an auto-encoder that maximizes the mutual information $I(x;g(x))$.

Please let me know if there are some missing mathematical technicalities and if the question needs further clarifications.


  • 3
    $\begingroup$ Please define your terms and notations. $\endgroup$ – Jan Johannsen Sep 11 '18 at 7:31

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.