adv-ml 0.0.2

A modular and easy-to-use framework of adversarial machine learning algorithms: https://en.m.wikipedia.org/wiki/Adversarial_machine_learning

adv-ml

Docs

See https://irad-zehavi.github.io/adv-ml/

Install

pip install adv_ml

How to use

How to Use

As an nbdev library, adv-ml supports import * (without importing unwanted symbols):

from adv_ml.all import *

Adversarial Examples

mnist = MNIST()
classifier = MLP(10)
learn = Learner(mnist.dls(), classifier, metrics=accuracy)
learn.fit(1)

epoch	train_loss	valid_loss	accuracy	time
0	0.160490	0.165644	0.954900	00:17

sub_dsets = mnist.valid.random_sub_dsets(64)
learn.show_results(shuffle=False, dl=sub_dsets.dl())

attack = InputOptimizer(classifier, LinfPGD(epsilon=.15), n_epochs=10)
perturbed_dsets = attack.perturb(sub_dsets)

epoch	train_loss	time
0	-3.627444	00:00
1	-6.452563	00:00
2	-7.652328	00:00
3	-8.258670	00:00
4	-8.617092	00:00
5	-8.851709	00:00
6	-9.014016	00:00
7	-9.130360	00:00
8	-9.216579	00:00
9	-9.281565	00:00

learn.show_results(shuffle=False, dl=TfmdDL(perturbed_dsets))