Great Article thanks, could we Speak?
I have posted a snippet of it linking back to this article on our website, is that ok with you?
All the best,
+44 7871 394 165
I’d say that information plane idea is great and examples with different learning algos behaving in the same way is interesting. Yet overall there are some controversies. E.g. in the way how mutual information should be computed for high-dimensional input or in no noticeable overfitting effects in experiments.