apt/index.html at main · CaptainE/apt · GitHub

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
<!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8">
  <meta name="description"
        content="Assessing Neural Network Robustness via Adversarial Pivotal Tuning">
  <meta name="keywords" content="Adversarial Pivotal Tuning, APT">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <title>Assessing Neural Network Robustness via Adversarial Pivotal Tuning</title>


  <link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
        rel="stylesheet">

  <link rel="stylesheet" href="./static/css/bulma.min.css">
  <link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
  <link rel="stylesheet" href="./static/css/bulma-slider.min.css">
  <link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
  <link rel="stylesheet"
        href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
  <link rel="stylesheet" href="./static/css/index.css">
  <link rel="icon" href="./static/images/favicon.svg">

  <script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
  <script defer src="./static/js/fontawesome.all.min.js"></script>
  <script src="./static/js/bulma-carousel.min.js"></script>
  <script src="./static/js/bulma-slider.min.js"></script>
  <script src="./static/js/index.js"></script>
</head>
<body>

<nav class="navbar" role="navigation" aria-label="main navigation">
  <div class="navbar-brand">
    <a role="button" class="navbar-burger" aria-label="menu" aria-expanded="false">
      <span aria-hidden="true"></span>
      <span aria-hidden="true"></span>
      <span aria-hidden="true"></span>
    </a>
  </div>
  <div class="navbar-menu">
    <div class="navbar-start" style="flex-grow: 1; justify-content: center;">
      <a class="navbar-item" href="https://keunhong.com">
      <span class="icon">
          <i class="fas fa-home"></i>
      </span>
      </a>

      <div class="navbar-item has-dropdown is-hoverable">
        <a class="navbar-link">
          More Research
        </a>
        <div class="navbar-dropdown">
          <a class="navbar-item" href="https://captaine.github.io/Searching-for-Structure-in-Unfalsifiable-Claims/">
            PAPYER
          </a>
        </div>
      </div>
    </div>

  </div>
</nav>


<section class="hero">
  <div class="hero-body">
    <div class="container is-max-desktop">
      <div class="columns is-centered">
        <div class="column has-text-centered">
          <h1 class="title is-1 publication-title">Assessing Neural Network Robustness via Adversarial Pivotal Tuning</h1>
          <div class="is-size-5 publication-authors">
            <span class="author-block">
              <a href="https://captaine.github.io/">Peter Ebert Christensen</a><sup>1</sup>,</span>
            <span class="author-block">
              <a href="https://vesteinn.is/">Vésteinn Snæbjarnarson</a><sup>1</sup>,</span>
            <span class="author-block">
              <a href="https://addtt.github.io/">Andrea Dittadi</a><sup>2</sup>,
            </span>
            <span class="author-block">
              <a href="https://sergebelongie.github.io/">Serge Belongie</a><sup>1</sup>,
            </span>
            <span class="author-block">
              <a href="https://sagiebenaim.github.io/">Sagie Benaim</a><sup>3</sup>
            </span>
          </div>

          <div class="is-size-5 publication-authors">
            <span class="author-block"><sup>1</sup>University of Copenhagen,</span>
            <span class="author-block"><sup>2</sup>Helmholtz AI</span>
            <span class="author-block"><sup>2</sup>Hebrew University of Jerusalem</span>
          </div>

          <div class="column has-text-centered">
            <div class="publication-links">
              <!-- PDF Link. -->
              <span class="link-block">
                <a href="https://arxiv.org/pdf/2211.09782"
                   class="external-link button is-normal is-rounded is-dark">
                  <span class="icon">
                      <i class="fas fa-file-pdf"></i>
                  </span>
                  <span>Paper</span>
                </a>
              </span>
              <span class="link-block">
                <a href="https://github.com/CaptainE/apt/tree/code"
                   class="external-link button is-normal is-rounded is-dark">
                  <span class="icon">
                    <i class="fab fa-github"></i>
                  </span>
                  <span>Code</span>
                </a>
              </span>
              <span class="link-block">
                <a href="https://arxiv.org/abs/2211.09782"
                   class="external-link button is-normal is-rounded is-dark">
                  <span class="icon">
                      <i class="ai ai-arxiv"></i>
                  </span>
                  <span>arXiv</span>
                </a>

            </div>

          </div>
        </div>
      </div>
    </div>
  </div>
</section>

<section class="hero teaser">
  <div class="container is-max-desktop">
    <div class="hero-body">
      <h2 class="subtitle has-text-centered">
        <span class="dnerf">APT</span> uses the full capacity of a pretrained generator to produce semantic adversarial manipulations
      </h2>
    </div>
  </div>
</section>


<section class="section">
    <div class="container is-max-desktop">
        <div class="columns is-centered has-text-centered">
            <div class="column is-full-width">

                <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
                <h2 class="title is-3">Visualizations</h2>
                <h3 class="title is-4">Overview of the Generated Manipulations</h3>
                <img srcset="teaser.png" alt="alt" width="1000px" height="563px" image-rendering: high-quality;>

                    <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
                <!-- Interpolating. -->

            </div>
        </div>
        <p>
          Row 1 shows the input images. Row 2 shows the images resulting from our manipulations. Row 3 (and 4)
shows the result of Dual manifold adversarial robustness, using pixel-space adversarial manipulations applied to StyleGAN-XL’s reconstructions. Row 4 shows the result
using latent space manipulates applied using StyleGAN-XL. Our method manipulates images in a non-trivial but class-preserving
manner, using the full capacity of a pretrained StyleGAN generator. For example, it removes the eye of the mantis (second column), changes
the type of race car (third column), changes the color of the crab tail (fifth column), removes the text in a spaceship (seventh column) and
removes some of the ropes (eighth column). All of these are class-preserving examples that fool a pretrained PRIME-ResNet50
classifier. In contrast, Dual manifold adversarial robustness either generates noisy and less realistic images (row 3) or images which differ significantly semantically and
which do not preserve the input class (row 4).
      </p>
    </div>
</section>


<section class="section">
  <div class="container is-max-desktop">
      <div class="columns is-centered has-text-centered">
          <div class="column is-full-width">

              <h2 class="title is-3">TLDR</h2>

          </div>
      </div>
      <p>
        We propose a framework for generation of photorealistic images that fool a classifier using automatic semantic manipulations
    </p>
  </div>
</section>


<section class="section">
  <div class="container is-max-desktop">

    <div class="column is-full-width  has-text-centered">

      <!-- Visual Effects. -->

        <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
        <h2 class="title is-3">The Adversarial Pivotal Tuning (APT) framework</h2>
        <img srcset="apt.png" alt="alt" width="1000px" height="563px" image-rendering: high-quality;>
    </div>

        <p>
            In the first step, we optimize a style code wp using standard latent optimization Lo, while keeping the generator G frozen.
            The loss is computed between the ground-truth image xgtr and the generated image xgen.
            In the second step, we freeze wp and finetune G (shown in red) using the three objectives; a
reconstruction objective Lrec, the projected GAN objective using the discriminator D, LP G, and our fooling objective LCE using the
classifier C. A ∗ is used to indicate a frozen component
          </p>
            <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
        <!-- Interpolating. -->

      <br>
      <br>
      <!--/ Visual Effects. -->


    <div class="column is-full-width  has-text-centered">

    <!-- Visual Effects. -->

        <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
        <h2 class="title is-3">Manipulations using different classifiers</h2>
        <img srcset="classifiers.png" alt="alt" width="1000px" height="563px" image-rendering: high-quality;>
    </div>

        <p>
            Top row shows input images. The middle row shows APT manipulations for a ResNet-50 classifier,
and the bottom row shows APT manipulations from a FAN-VIT classifier. The leftmost image of a dog and the subsequent images including
the image of a butterfly and column 7 (Fluffy dog) show similar manipulations for both classifiers, column 5-6 shows texture and spatial
manipulations, the last column showcase a fooling image without a clear APT manipulati
        </p>
            <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
        <!-- Interpolating. -->

    <!--/ Visual Effects. -->
    <br>
    <br>


    <div class="column is-full-width  has-text-centered">

    <!-- Visual Effects. -->

        <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
        <h2 class="title is-3">Transferability of APT generated samples</h2>
        <img srcset="Transferability.png" alt="alt" width="1000px" height="563px" image-rendering: high-quality;>
      </div>
        <p>
             For the ImageNet-1k validation set, we consider samples generated to fool a PRIME-Resnet50 (PRIME) and a FAN-VIT (FAN) pretrained classifier. We then test the accuracy (Acc) and mean softmax probability
of the labelled class (Conf) on those samples. The left column indicates the classifier on which we tested the accuracy of real or generatedsamples. ∗ indicates the accuracy and confidence of samples generated and tested using the same classifier.
        </p>
            <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
        <!-- Interpolating. -->

    <!--/ Visual Effects. -->
    <br>
    <br>


    <div class="column is-full-width  has-text-centered">

      <!-- Visual Effects. -->

          <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
          <h2 class="title is-3">Attack Success Rate (ASR) for APT, SSAH and PGD</h2>
          <img srcset="asr.png" alt="alt" width="640px" height="420px" image-rendering: high-quality;>
        </div>
          <p>
               We investigate the attack success rate for our method APT against more traditional but imperceptible attacks such as PGD and SSAH.
               Additionally we also test how well these generated images can fool other classifiers that was not subject to an attack.
          </p>
          <div class="column is-full-width  has-text-centered">

          <p>
            An example of these attacks are illustrated here
           </p>
          <img srcset="eagle.png" alt="alt" width="640px" height="420px" image-rendering: high-quality;>
        </div>

              <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
          <!-- Interpolating. -->

      <!--/ Visual Effects. -->
      <br>
      <br>


    <!-- Visual Effects. -->
    <div class="column is-full-width  has-text-centered">
    <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
    <h2 class="title is-3">Average accuracy and confidence on APT samples using PRIME-ResNet50 before and after fine-tuning.</h2>
    <img srcset="finetune.png" alt="alt" width="500px" height="263px" image-rendering: high-quality;>
  </div>
    <p>
         We investigate the effect of finetuning a PRIME-ResNet50 model on our generated fooling images usng APT, PGD and SSAG.
         We find that the accuracy to correcly predict the original class increases after finetuning for each indidual attack
         but also when combining all attacks.
    </p>
    <br>
    <br>


    <!-- Visual Effects. -->

    <hr style="height:2px;border-width:0;color:gray;background-color:gray;margin-top:-50px;">
    <h2 class="title is-3">Acknowledgement</h2>
    <p>
      This research was supported by the Pioneer Centre for AI, DNRF grant number P1.
    </p>
        <!-- <embed src="taxon_labels_numbers.png" width="1000px" height="563px" /> -->
    <!-- Interpolating. -->


<!--/ Visual Effects. -->
<br>
<br>


  </div>
</section>


<section class="section" id="BibTeX">
  <div class="container is-max-desktop content">
    <h2 class="title">BibTeX</h2>
    <pre><code>@article{christensen2022apt
    author    = {Christensen, Peter Ebert and Snæbjarnarson, Vésteinn and Dittadi, Andrea and Belongie, Serge and Benaim, Sagie},
    title     = {Assessing Neural Network Robustness via Adversarial Pivotal Tuning},
    journal   = {arxiv},
    year      = {2022},
  }
</code></pre>
  </div>
</section>


<footer class="footer">
  <div class="container">
    <div class="content has-text-centered">
      <a class="icon-link"
         href="https://homes.cs.washington.edu/~kpar/nerfies/videos/nerfies_paper.pdf">
        <i class="fas fa-file-pdf"></i>
      </a>
      <a class="icon-link" href="https://github.com/keunhong" class="external-link" disabled>
        <i class="fab fa-github"></i>
      </a>
    </div>
    <div class="columns is-centered">
      <div class="column is-8">
        <div class="content">
          <p>
            This website is licensed under a <a rel="license"
                                                href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
            Commons Attribution-ShareAlike 4.0 International License</a>.
          </p>
        </div>
      </div>
    </div>
  </div>
</footer>

</body>
</html>