News for hackers. Stuff that matters.
Re:Understanding deep learning requires rethinking generaliz... by anonymousDan on Friday March 05, 2021 @07:19AM (#26356008)
But isn't that shown by the fact it takes training longer to converge in their experiments with random labels vs real labels?
Re:Understanding deep learning requires rethinking generaliz... by sdenton4 on Friday March 05, 2021 @11:00AM (#26358275)
Yeah, not a bad idea; for any given problem, just scramble the labels and compare training time...

For my pet problem, there's quite different performance across classes. It would be tougher to do the comparison for a subset of classes, but could still probably design some kind of experiment.