News for hackers. Stuff that matters.

A comment on Computer Vision Research: The deep “depression”. Parent comment

Re:Computer Vision Research: The deep “depression” by argonaut on Thursday June 09, 2016 @12:42PM (#11870538)
The chicken and egg problem is that we design neural nets in order for Gabor filters to show up. If you used a different neural net architecture choice, they wouldn't show up. So the presence of Gabor filters indicating optimality is sort of begging the question.
[ Parent | on HN ]
Re:Computer Vision Research: The deep “depression” by hacker42 on Thursday June 09, 2016 @01:37PM (#11870934)
As far as I know, Gabor filters show up in all kinds of multi-layered models, even in unsupervised ones. Gabor filters are quite canonical.
[ Parent | on HN ]
Re:Computer Vision Research: The deep “depression” by trhway on Thursday June 09, 2016 @12:55PM (#11870616)
CNN beat other architectures in images and Gabor emerges in CNN. That is a pretty good candidate for even global optimaliy.
[ Parent | on HN ]
Re:Computer Vision Research: The deep “depression” by argonaut on Thursday June 09, 2016 @01:34PM (#11870910)
The fact that X works and A, B, and C don't work, does not imply A is globally optimal.
[ Parent | on HN ]