NIPS2017_adv_challenge_defense icon indicating copy to clipboard operation
NIPS2017_adv_challenge_defense copied to clipboard

Published 20 hours ago • verified publisher icon cihangxie

Still work on MNIST?

Open HaoerSlayer opened this issue 4 years ago • 1 comments

Hi,

I was trying this defense with a model with two conv layers trained on MNIST. However,it seems useless, which improves the acc under fgsm from 4.2% to 16.04%. Because of the restriction of my calculation power, I cant do with the images with size of 299*299, so Im not sure if this poor improve is caused by my wrong implementation or the diff of data. Could you give me some suggestion?

HaoerSlayer avatar Jun 06 '20 00:06 HaoerSlayer

I have not tried any experiments on MNIST.

But a precondition for allowing this defense to work is that your classier should work well when applying resizing operation to clean images. e.g., if the accuracy of your classifier on clean inputs is 99%, then the accuracy of your classifier on RESIZED clean inputs should also be ~99%.

You do not need to resize the MNIST images to 299 (kind of too large). Just resizing them between the range [30, 40] may work.

cihangxie avatar Jun 10 '20 02:06 cihangxie