View Single Post
Old 01-25-2019, 06:46 AM
Yew Lee Yew Lee is offline
Junior Member
Join Date: Jan 2019
Posts: 3
Default Re: Paradox in VC dimension

B's model cannot be better, because A and B used the same model.

If you mean B happened to choose a simpler network than A did (e.g. less layers), then since A is more educated clearly he would know how to put weights regularization, dropout,... to avoid overfitting.
I think what PCdimension meant is that since researcher (A)'s hypothesis set contains 2 different models, the VC dimension will be larger. Therefore although both researchers submitted the same model, the out of sample error bound for researcher (B) is smaller than that of researcher (A).
Reply With Quote