-
Notifications
You must be signed in to change notification settings - Fork 1
/
model_utils.py
400 lines (331 loc) · 13.1 KB
/
model_utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
from __future__ import absolute_import
from __future__ import division
from __future__ import unicode_literals
import numpy as np
import tensorflow as tf
from tensorflow.python.platform import flags
import logging
import sys
import pickle
from sklearn import tree
from sklearn.preprocessing import normalize
from cleverhans.utils_mnist import data_mnist
from cleverhans.utils_tf import model_train, model_eval
from cleverhans.attacks import FastGradientMethod
from cleverhans_tutorials.tutorial_models import make_basic_cnn, MLP
from cleverhans.utils import AccuracyReport, set_log_level
from cleverhans_tutorials.tutorial_models import Layer, Flatten, Linear, ReLU, Softmax
class Sigmoid(Layer):
def __init__(self):
pass
def set_input_shape(self, shape):
self.input_shape = shape
self.output_shape = shape
def get_output_shape(self):
return self.output_shape
def fprop(self, x):
return tf.sigmoid(x)
def basic_mlp(nb_classes=2, input_shape=None):
layers = [Flatten(),
Linear(20),
Sigmoid(),
Linear(10),
Sigmoid(),
Flatten(),
Linear(nb_classes),
Softmax()
]
model = MLP(layers, input_shape)
return model
def abalone_mlp(nb_classes=2, input_shape=None):
layers = [Flatten(),
Linear(10),
Sigmoid(),
Linear(10),
Sigmoid(),
Flatten(),
Linear(nb_classes),
Softmax()
]
model = MLP(layers, input_shape)
return model
def make_model(task):
if task == 'mnist':
model = make_basic_cnn(nb_classes=2)
elif task == 'abalone':
model = abalone_mlp(nb_classes=2, input_shape=[None, 7, 1, 1])
else:
model = basic_mlp(nb_classes=2, input_shape=[None, 2, 1, 1])
return model
class neural_net_classifier():
def __init__(self, X=None, y=None, model=None, task=None):
self.X = X
self.y = y
self.model = model
self.task = task
def fit(self, X, y, target_model=None, model=None, train_params=None, shape=None, num_rounds=3):
self.X = X
self.y = y
self.train_params = train_params
self.shape = shape
self.model = model
#num_rounds=3
eps_aug = 2
train_params = self.train_params
nb_epochs= train_params['nb_epochs']
batch_size=train_params['batch_size']
learning_rate=train_params['learning_rate']
for i in range(num_rounds):
# Set TF random seed to improve reproducibility
tf.set_random_seed(1234)
tf.reset_default_graph()
# Create TF session
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.3)
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
# Define input TF placeholder
x = tf.placeholder(tf.float32, shape=self.shape)
y = tf.placeholder(tf.float32, shape=(None, 2))
rng = np.random.RandomState([2017, 8, 30])
task = self.task
model = make_model(task)
preds = model.get_probs(x)
def evaluate():
pass
X_train = self.X
Y_train = self.y
shape_original = X_train.shape
X_train = np.reshape(X_train, [len(X_train)]+self.shape[1:])
Y_train = np.array([[1,0] if i==1 else [0,1] for i in Y_train])
model_train(sess, x, y, preds, X_train, Y_train, evaluate=evaluate,
args=train_params, rng=rng)
eval_params = {'batch_size': batch_size}
pred_results = sess.run(preds, feed_dict={x:X_train})
fgsm_params = {'eps': eps_aug,
#'clip_min': 0.0,
#'clip_max': 1.0,
'ord': 2}
fgsm = FastGradientMethod(model, sess=sess)
adv_x = fgsm.generate(x, **fgsm_params)
preds_adv = model.get_probs(adv_x)
X_adv = sess.run(adv_x, feed_dict={x: X_train})
X_adv = np.reshape(X_adv, shape_original)
print X_adv.shape
X_adv = np.nan_to_num(np.clip(X_adv, -100, 100))
y_adv = target_model.predict(X_adv)
self.y = np.concatenate((self.y, y_adv), axis=0)
self.X = np.concatenate((self.X, X_adv), axis=0)
sess.close()
del sess
def predict(self, X_test):
train_params = self.train_params
nb_epochs= train_params['nb_epochs']
batch_size=train_params['batch_size']
learning_rate=train_params['learning_rate']
tf.set_random_seed(1234)
tf.reset_default_graph()
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.3)
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
# Define input TF placeholder
x = tf.placeholder(tf.float32, shape=self.shape)
y = tf.placeholder(tf.float32, shape=(None, 2))
rng = np.random.RandomState([2017, 8, 30])
task = self.task
model = make_model(task)
preds = model.get_probs(x)
def evaluate():
pass
X_train = self.X
Y_train = self.y
X_train = np.reshape(X_train, [len(X_train)]+self.shape[1:])
X_test = np.reshape(X_test, [len(X_test)]+self.shape[1:])
Y_train = np.array([[1,0] if i==1 else [0,1] for i in Y_train])
model_train(sess, x, y, preds, X_train, Y_train, evaluate=evaluate,
args=train_params, rng=rng)
eval_params = {'batch_size': batch_size}
pred_results = sess.run(preds, feed_dict={x:X_test})
sess.close()
del sess
pred_results = np.array([1 if a[0]>a[1] else -1 for a in pred_results])
return pred_results
def generate_adv(self, X_test, eps, mask=None):
train_params = self.train_params
nb_epochs= train_params['nb_epochs']
batch_size=train_params['batch_size']
learning_rate=train_params['learning_rate']
tf.set_random_seed(1234)
tf.reset_default_graph()
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.3)
sess = tf.Session(config=tf.ConfigProto(gpu_options=gpu_options))
# Define input TF placeholder
x = tf.placeholder(tf.float32, shape=self.shape)
y = tf.placeholder(tf.float32, shape=(None, 2))
rng = np.random.RandomState([2017, 8, 30])
task = self.task
model = make_model(task)
preds = model.get_probs(x)
def evaluate():
pass
X_train = self.X
Y_train = self.y
X_train = np.reshape(X_train, [len(X_train)]+self.shape[1:])
shape_original = X_test.shape
X_test_original = np.copy(X_test)
X_test = np.reshape(X_test, [len(X_test)]+self.shape[1:])
Y_train = np.array([[1,0] if i==1 else [0,1] for i in Y_train])
model_train(sess, x, y, preds, X_train, Y_train, evaluate=evaluate,
args=train_params, rng=rng)
eval_params = {'batch_size': batch_size}
pred_results = sess.run(preds, feed_dict={x:X_test})
fgsm_params = {'eps': eps,
'ord': 2}
fgsm = FastGradientMethod(model, sess=sess)
adv_x = fgsm.generate(x, **fgsm_params)
preds_adv = model.get_probs(adv_x)
X_adv = sess.run(adv_x, feed_dict={x: X_test})
X_adv = np.reshape(X_adv, shape_original)
for i in range(len(mask)):
if mask[i]:
pass
else:
X_adv[i] = X_test_original[i]
sess.close()
del sess
return X_adv
###########################
### kernel classifier
class Kernel_Classifier():
def __init__(self, X=None, y=None, c=None, eps=None, num_rounds=None):
self.X = X
self.y = y
self.c = c
def augment(self, eps, num_rounds):
[X, y] = [self.X, self.y]
for i in range(num_rounds):
mask_pred = [True for x in X]
X_adv = self.generate_adv(X, y, eps, mask_pred)
X = np.concatenate([X, X_adv], axis=0)
y = np.concatenate([y, y])
self.X = X
self.y = y
def fit(self, X, y):
self.X = X
self.y = y
def predict(self, X_test):
X_train_kernel = self.X
y_train_kernel = self.y
c = self.c
n_test = len(X_test)
n_train_kernel = len(X_train_kernel)
d = len(X_test[0])
y_pred = np.zeros([n_test, 1])
for i in range(n_test):
x = X_test[i].reshape(1,d)
#y = y_test[i]
x = np.repeat(x, n_train_kernel, axis=0)
delta = (X_train_kernel - x)
dist = np.zeros([n_train_kernel, 1])
e = np.zeros([n_train_kernel, 1])
for k in range(n_train_kernel):
dist[k] = np.linalg.norm(x[k]-X_train_kernel[k])
e[k] = np.exp(-dist[k]*1.0/c)
mask = (y_train_kernel==1).reshape(n_train_kernel,1)
total = sum(e)
p_pos = sum(np.multiply(e, mask))
p_neg = total - p_pos
if p_pos > p_neg:
y_pred[i] = 1
else:
y_pred[i] = -1
return y_pred
def generate_adv(self, X_test, y_test, eps, mask_pred):
X_train_kernel = self.X
y_train_kernel = self.y
c = self.c
n_test = len(X_test)
n_train_kernel = len(X_train_kernel)
d = len(X_test[0])
adv = np.zeros([n_test, d])
for i in range(n_test):
x = X_test[i].reshape(1,d)
y = y_test[i]
x = np.repeat(x, n_train_kernel, axis=0)
delta = (X_train_kernel - x)*2
dist = np.zeros([n_train_kernel, 1])
e = np.zeros([n_train_kernel, 1])
for k in range(n_train_kernel):
dist[k] = np.linalg.norm(x[k]-X_train_kernel[k])
e[k] = np.exp(-dist[k]*1.0/c)
if y==1:
mask = (y_train_kernel==1)
else:
#mask = 1-y_train_kernel
mask = 1-(y_train_kernel==1)
mask = mask.reshape([len(mask),1])
g_same = np.multiply(e, mask)
delta_same = np.multiply(delta, mask)
sum_e = sum(e)
sum_g_same = sum(g_same)
g_2 = sum_e**2
t_same = np.repeat(g_same, d, axis=1)
df_g = np.multiply(t_same, delta_same) * sum_e
df_g = np.sum(df_g, axis=0).reshape(1, d)
t = np.repeat(e, d, axis=1)
dg_f = np.multiply(t, delta) * sum_g_same
dg_f = np.sum(dg_f, axis=0).reshape(1,d)
deriv = (df_g - dg_f)*1.0
deriv = normalize(deriv)
X_new = x[0]-eps*deriv
if mask_pred[i]:
adv[i] = X_new
else:
adv[i] = X_test[i]
return adv
# DT attacks
class decisionTreeNode:
def __init__(self, node_id=None, input_component=None, threshold=None,left=None,right=None,output=[], parent=None):
self.node_id = node_id
self.input_component = input_component
self.threshold = threshold
self.left = left
self.right = right
self.output = output
self.parent = parent
def tree_parser(clf):
t = clf.tree_
n_nodes = t.node_count
children_left = t.children_left
children_right = t.children_right
feature = t.feature
threshold = t.threshold
values = t.__getstate__()['values']
t_dict = {}
stack = [(0, None)] # seed is the root node id and its parent depth
while len(stack) > 0:
node_id, parent_id = stack.pop()
# If we have a test node
if (children_left[node_id] != children_right[node_id]):
left = children_left[node_id]
right = children_right[node_id]
input_component = feature[node_id]
thres = threshold[node_id]
output = values[node_id]
stack.append((left, node_id))
stack.append((right, node_id))
if parent_id:
t_dict[str(node_id)] = decisionTreeNode(str(node_id), input_component,
thres, str(left), str(right), output, str(parent_id))
else:
t_dict[str(node_id)] = decisionTreeNode(str(node_id), input_component,
thres, str(left), str(right), output, None)
else:
input_component = feature[node_id]
thres = threshold[node_id]
output = values[node_id]
t_dict[str(node_id)] = decisionTreeNode(str(node_id), input_component,
thres, None, None, output, str(parent_id))
return t_dict
def prepare_tree(X, y, max_depth):
clf = tree.DecisionTreeClassifier(criterion="gini", max_depth=max_depth)
clf.fit(X, y)
dt = tree_parser(clf)
return clf, dt