def fig11(tr_data, tr_labels, te_data, te_labels, tr_files):
"""
Tests the vaccination defense against the Benign Random Noise (BRN)
attack seeded by results of our mimicry attack against itself and
original, unmodified data. Performs 5 trials.
"""
mal_tr_ind = [i for i, l in enumerate(tr_labels) if l == 1]
ben_tr_ind = [i for i, l in enumerate(tr_labels) if l == 0]
mim_data, mim_labels = common.get_FTC_mimicry()
TRIALS = 5
print "\n{:>6}{:>15}{:>15}".format("%", "ORIGINAL", "OUR MIMICRY")
pool = multiprocessing.Pool(processes=None)
scores = []
for subset in (0, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1):
acc = [0.0, 0.0]
sys.stdout.write("{:>6.2f}".format(subset * 100))
for _ in range(TRIALS):
tr_mod = tr_data.copy()
# Subsample malicious training files for attack
wolf_ind = random.sample(mal_tr_ind, int(round(subset * len(mal_tr_ind))))
# Mimic random benign files using the sampled files
pargs = [(tr_data[random.choice(ben_tr_ind)], tr_files[w_id], w_id) for w_id in wolf_ind]
for mimic, w_id in pool.imap(mimicry_wrap, pargs):
tr_mod[w_id] = mimic
# Evaluate the classifier on both clean test data and mimicry data
res = common.evaluate_classifier(tr_mod, tr_labels, [te_data, mim_data], [te_labels, mim_labels])
acc = [old + new for old, new in zip(acc, res)]
acc = [acc[0] / TRIALS, acc[1] / TRIALS]
print "{:>15.3f}{:>15.3f}".format(acc[0], acc[1])
scores.append(tuple(acc))
return scores
def fig10(data, labels):
'''
Reproduction of results published in Table 12 of "Malicious PDF
Detection Using Metadata and Structural Features" by Charles Smutz
and Angelos Stavrou, ACSAC 2012.
'''
ben_means, ben_devs = common.get_benign_mean_stddev(data, labels)
mim_data, mim_labels = common.get_FTC_mimicry()
TRIALS = 5
nCV = 10
subsets = [0, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1]
pool = multiprocessing.Pool(processes=None)
pool_args = [(data, labels, mim_data, mim_labels, ben_means, ben_devs,
subset, TRIALS, nCV) for subset in subsets]
print '\n % {:>15}{:>15}{:>15}'.format('ORIGINAL', 'MIMICRY',
'OUR MIMICRY'),
norm = TRIALS * nCV
res = []
for accs, subset in pool.imap(perturbate_CV_parallel, pool_args):
print '\n{:>6.2f}'.format(subset * 100),
for acc in accs:
sys.stdout.write('{:>15.3f}'.format(acc / norm))
res.append(tuple([acc / norm for acc in accs]))
return res
def fig11(tr_data, tr_labels, te_data, te_labels, tr_files):
'''
Tests the vaccination defense against the Benign Random Noise (BRN)
attack seeded by results of our mimicry attack against itself and
original, unmodified data. Performs 5 trials.
'''
mal_tr_ind = [i for i, l in enumerate(tr_labels) if l == 1]
ben_tr_ind = [i for i, l in enumerate(tr_labels) if l == 0]
mim_data, mim_labels = common.get_FTC_mimicry()
TRIALS = 5
print '\n{:>6}{:>15}{:>15}'.format('%', 'ORIGINAL', 'OUR MIMICRY')
pool = multiprocessing.Pool(processes=None)
scores = []
for subset in (0, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1):
acc = [0.0, 0.0]
sys.stdout.write('{:>6.2f}'.format(subset * 100))
for _ in range(TRIALS):
tr_mod = tr_data.copy()
# Subsample malicious training files for attack
wolf_ind = random.sample(mal_tr_ind,
int(round(subset * len(mal_tr_ind))))
# Mimic random benign files using the sampled files
pargs = [(tr_data[random.choice(ben_tr_ind)], tr_files[w_id], w_id)
for w_id in wolf_ind]
for mimic, w_id in pool.imap(mimicry_wrap, pargs):
tr_mod[w_id] = mimic
# Evaluate the classifier on both clean test data and mimicry data
res = common.evaluate_classifier(tr_mod, tr_labels,
[te_data, mim_data],
[te_labels, mim_labels])
acc = [old + new for old, new in zip(acc, res)]
acc = [acc[0] / TRIALS, acc[1] / TRIALS]
print '{:>15.3f}{:>15.3f}'.format(acc[0], acc[1])
scores.append(tuple(acc))
return scores
def fig10(data, labels):
"""
Reproduction of results published in Table 12 of "Malicious PDF
Detection Using Metadata and Structural Features" by Charles Smutz
and Angelos Stavrou, ACSAC 2012.
"""
ben_means, ben_devs = common.get_benign_mean_stddev(data, labels)
mim_data, mim_labels = common.get_FTC_mimicry()
TRIALS = 5
nCV = 10
subsets = [0, 0.0005, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1]
pool = multiprocessing.Pool(processes=None)
pool_args = [(data, labels, mim_data, mim_labels, ben_means, ben_devs, subset, TRIALS, nCV) for subset in subsets]
print "\n % {:>15}{:>15}{:>15}".format("ORIGINAL", "MIMICRY", "OUR MIMICRY"),
norm = TRIALS * nCV
res = []
for accs, subset in pool.imap(perturbate_CV_parallel, pool_args):
print "\n{:>6.2f}".format(subset * 100),
for acc in accs:
sys.stdout.write("{:>15.3f}".format(acc / norm))
res.append(tuple([acc / norm for acc in accs]))
return res