-
Notifications
You must be signed in to change notification settings - Fork 2
Expand file tree
/
Copy pathindex.html
More file actions
411 lines (378 loc) · 20.7 KB
/
index.html
File metadata and controls
411 lines (378 loc) · 20.7 KB
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
401
402
403
404
405
406
407
408
409
410
411
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8" />
<meta name="description" content="DetoxAI is a Python package for debiasing neural networks in image classification tasks." />
<meta property="og:title" content="DetoxAI: a Python package for debiasing neural networks" />
<meta property="og:description" content="DetoxAI is a Python package for debiasing neural networks in image classification tasks." />
<meta property="og:url" content="https://github.com/DetoxAI/detoxai" />
<meta name="twitter:title" content="DetoxAI: a Python package for debiasing neural networks" />
<meta name="twitter:description" content="DetoxAI is a Python package for debiasing neural networks in image classification tasks." />
<meta name="twitter:card" content="summary" />
<meta name="keywords" content="artificial intelligence, machine learning, neural networks, image classification, fairness, debiasing" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<title>DetoxAI: a Python Package for debiasing neural networks</title>
<link rel="icon" type="image/x-icon" href="static/images/favicon.ico" />
<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="manifest" href="site.webmanifest" />
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.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>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<div class="publication-title">
<h1 class="title is-1" style="font-size: 5rem;">
<img src="static/images/android-chrome-512x512.png" alt="DetoxAI Logo" style="height: 1.2em; margin-right: 10px; vertical-align: middle; padding: 0; display: inline-block; position: relative; top: -0.1em; border-radius: 10px;">
DetoxAI
</h1>
<h2 class="subtitle is-3" style="font-weight: normal; margin-top: 0;">
A Python package for debiasing neural networks
</h2>
</div>
<div class="is-size-5 publication-authors" style="margin-top: 1rem;">
<span class="author-block">
<a href="https://ignacystepka.com" target="_blank">Ignacy Stepka</a><sup>*</sup>
</span>
<span class="author-block">
<a href="https://lukaszsztukiewicz.com" target="_blank">Lukasz Sztukiewicz</a><sup>*</sup>
</span>
<span class="author-block">
<a href="https://michal-wilinski.com" target="_blank">Michal Wilinski</a><sup>*</sup>
</span>
<span class="author-block">
<a href="https://www.cs.put.poznan.pl/jstefanowski/" target="_blank">Jerzy Stefanowski</a>
</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">
<span style="display: inline-flex; align-items: center; justify-content: center; margin-bottom: 8px;">
<img src="static/images/put_logo.png" alt="PUT Logo" style="height: 1.5em; margin-right: 10px;">
<span style="padding-top: 3px; padding-bottom: 3px;">Poznan University of Technology</span>
</span>
<br>
</span>
</div>
<span class="eql-cntrb is-size-6">
<small><sup>*</sup>Indicates Equal Contribution</small>
</span>
<div class="publication-venue" style="margin: 1rem auto;">
Published at ECML PKDD 2025, Demo Track
</div>
<div class="publication-links" style="margin-top: 1rem; margin-bottom: 1rem;">
<span class="link-block">
<a href="https://github.com/DetoxAI/detoxai" target="_blank" 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://detoxai.readthedocs.io" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-book-open"></i>
</span>
<span>Documentation</span>
</a>
</span>
<span class="link-block">
<a href="https://youtu.be/3TPgEoQnlbA?si=-pWr3v_-kVhCot_8" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-video"></i>
</span>
<span>Video</span>
</a>
</span>
<span class="link-block">
<a href="static/pdfs/paper.pdf" target="_blank" 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://link.springer.com/chapter/10.1007/978-3-032-06129-4_39" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-springer"></i>
</span>
<span>Springer</span>
</a>
</span>
</div>
<div style="margin-top: 2rem;">
<h2 class="title is-4 has-text-centered">How DetoxAI Works</h2>
<figure class="image">
<img src="static/images/visual_abstract.webp" alt="DetoxAI Features" style="max-width: 95%; margin: 0 auto; border-radius: 8px;" />
<figcaption class="has-text-centered" style="margin-top: 0.75rem; font-size: 0.9rem; color: #666;">
DetoxAI transforms biased neural networks into fair and balanced models with minimal code integration.
</figcaption>
</figure>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section" style="padding-top: 0;">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column is-four-fifths">
<h2 class="title is-3 has-text-centered">Key Features</h2>
<div class="columns is-multiline">
<div class="column is-half">
<div class="card" style="height: 100%; display: flex; flex-direction: column;">
<div class="card-content" style="display: flex; flex-direction: column; flex-grow: 1;">
<p class="title is-4">
<span class="icon" style="color: #01628b;">
<i class="fas fa-tools"></i>
</span>
Multiple Debiasing Methods
</p>
<div class="content" style="flex-grow: 1; display: flex; flex-direction: column;">
<div>
<p>DetoxAI includes several state-of-the-art debiasing techniques:</p>
<ul>
<li>
ClArC family
<a href="https://www.sciencedirect.com/science/article/pii/S1566253521001573" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
<li>
Zhang et al.
<a href="https://arxiv.org/abs/1801.07593" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
<li>
Savani et al.
<a href="https://arxiv.org/abs/2006.08564" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
<li>
Belrose et al.
<a href="https://arxiv.org/abs/2306.03819" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
</ul>
</div>
<div class="mt-4 has-text-right" style="margin-top: 2rem !important;">
<a href="https://detoxai.readthedocs.io/en/latest/tutorials.adding_a_method.html" target="_blank">Learn how to add your own method →</a>
</div>
</div>
</div>
</div>
</div>
<div class="column is-half">
<div class="card" style="height: 100%; display: flex; flex-direction: column;">
<div class="card-content" style="display: flex; flex-direction: column; flex-grow: 1;">
<p class="title is-4">
<span class="icon" style="color: #01628b;">
<i class="fas fa-database"></i>
</span>
Dataset Integration
</p>
<div class="content" style="flex-grow: 1; display: flex; flex-direction: column;">
<div>
<p>Works with popular datasets out of the box:</p>
<ul>
<li>
CelebA
<a href="https://mmlab.ie.cuhk.edu.hk/projects/CelebA.html" target="_blank" title="Visit dataset page">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-external-link-alt"></i></span>
</a>
</li>
<li>
FairFace
<a href="https://arxiv.org/abs/1908.04913" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
<li>
CIFAR-10/100
<a href="https://www.cs.toronto.edu/~kriz/cifar.html" target="_blank" title="Visit dataset page">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-external-link-alt"></i></span>
</a>
</li>
<li>
Caltech101
<a href="https://data.caltech.edu/records/mzrjq-6wc02" target="_blank" title="Visit dataset page">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-external-link-alt"></i></span>
</a>
</li>
</ul>
</div>
<div class="mt-4 has-text-right" style="margin-top: 2rem !important;">
<a href="https://detoxai.readthedocs.io/en/latest/tutorials.dataset.html" target="_blank">Learn how to use datasets →</a>
</div>
</div>
</div>
</div>
</div>
<div class="column is-half">
<div class="card" style="height: 100%; display: flex; flex-direction: column;">
<div class="card-content" style="display: flex; flex-direction: column; flex-grow: 1;">
<p class="title is-4">
<span class="icon" style="color: #01628b;">
<i class="fas fa-chart-line"></i>
</span>
Visualization Tools
</p>
<div class="content" style="flex-grow: 1; display: flex; flex-direction: column;">
<div>
<p>Powerful tools to understand and visualize bias:</p>
<ul>
<li>
Saliency maps with Layer-wise Relevance Propagation
<a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0130140" target="_blank" title="View paper">
<span class="icon is-small" style="color: #01628b;"><i class="fas fa-file-alt"></i></span>
</a>
</li>
<li>Side-by-side comparisons of original vs. debiased models</li>
<li>Aggregate visualizations to track focus shift</li>
</ul>
</div>
<div class="mt-4 has-text-right" style="margin-top: 2rem !important;">
<a href="https://detoxai.readthedocs.io/en/latest/detoxai.visualization.html" target="_blank">Explore visualization tools →</a>
</div>
</div>
</div>
</div>
</div>
<div class="column is-half">
<div class="card" style="height: 100%; display: flex; flex-direction: column;">
<div class="card-content" style="display: flex; flex-direction: column; height: 100%;">
<p class="title is-4">
<span class="icon" style="color: #01628b;">
<i class="fas fa-laptop-code"></i>
</span>
Simple API
</p>
<div class="content" style="display: flex; flex-direction: column; flex: 1;">
<div>
<p>DetoxAI is designed to be easy to use:</p>
<ul>
<li>Works with existing PyTorch models</li>
<li>Integrates with standard dataloaders</li>
<li>Single function call to apply multiple methods</li>
</ul>
</div>
<div style="margin-top: auto; padding-top: 2rem; text-align: right;">
<a href="https://detoxai.readthedocs.io/en/latest/examples/example.html" target="_blank">See a complete example →</a>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<h2 class="title is-3 has-text-centered">Quick Start</h2>
<div class="code-snippet">
<div class="code-header" style="background-color: #1e1e1e; color: #d4d4d4; padding: 8px 15px; border-top-left-radius: 6px; border-top-right-radius: 6px; display: flex; align-items: center;">
<span style="margin-right: 10px;"><i class="fas fa-code"></i></span>
<span>Code Example</span>
</div>
<pre class="language-python" style="margin-top: 0; border-top-left-radius: 0; border-top-right-radius: 0;"><code class="language-python"><span style="color: #88846f;"># DetoxAI is available on PyPI, and can be installed by running the following command:</span>
<span style="color: #005eff;">pip install detoxai</span>
<span style="color: #88846f;"># DetoxAI Quickstart Example</span>
<span style="color: #ff3535;">import</span> <span style="color: #005eff;">detoxai</span>
<span style="color: #88846f;"># Load your pretrained model</span>
model = ...
<span style="color: #88846f;"># Prepare a dataloader yielding (input, label, protected attribute)</span>
dataloader = ...
<span style="color: #88846f;"># Debias the model using DetoxAI</span>
corrected = detoxai.<span style="color: #ff3535;">debias</span>(model, dataloader)
<span style="color: #88846f;"># Access metrics from the SAVANIAFT debiasing method</span>
metrics = corrected[<span style="color: #005eff;">"SAVANIAFT"</span>].<span style="color: #ff3535;">get_all_metrics</span>()
<span style="color: #88846f;"># Retrieve the debiased model</span>
model = corrected[<span style="color: #005eff;">"SAVANIAFT"</span>].<span style="color: #ff3535;">get_model</span>()
</code></pre>
</div>
<h2 class="subtitle has-text-centered" style="margin-top: 1.5rem;">
DetoxAI provides a simple API for debiasing neural networks in image classification tasks.
</h2>
</div>
</div>
</section>
<section class="hero is-small">
<div class="hero-body">
<div class="container">
<div class="columns is-centered">
<div class="column">
<figure class="image">
<img src="static/images/side-by-side.webp" alt="Side by side comparison" />
<h2 class="subtitle has-text-centered">
Saliency maps obtained with Layer-wise Relevance Propagation (LRP) showing where the model focuses when classifying images.
</h2>
</figure>
</div>
<div class="column">
<figure class="image">
<img src="static/images/aggregate.webp" alt="Aggregate results" />
<h2 class="subtitle has-text-centered">
Aggregate visualizations (average photos) that we used in our work to show the effect of debiasing.
</h2>
</figure>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>
@inproceedings{stkepka2025detoxai,
title={DetoxAI: a Python Toolkit for Debiasing Deep Learning Models in Computer Vision},
author={St{\k{e}}pka, Ignacy and Sztukiewicz, Lukasz and Wili{\'n}ski, Micha{\l} and Stefanowski, Jerzy},
booktitle={Joint European Conference on Machine Learning and Knowledge Discovery in Databases},
pages={502--505},
year={2025},
organization={Springer}
}
</code></pre>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This page was built using the
<a href="https://github.com/eliahuhorwitz/Academic-project-page-template" target="_blank">Academic Project Page Template</a>
which was adopted from the <a href="https://nerfies.github.io" target="_blank">Nerfies</a> project page.
<br />
This website is licensed under a
<a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>