-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
488 lines (439 loc) · 22.1 KB
/
index.html
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
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
<!DOCTYPE HTML>
<!--
Dimension by HTML5 UP
html5up.net | @ajlkn
Free for personal and commercial use under the CCA 3.0 license (html5up.net/license)
-->
<html>
<head>
<title>6PAC</title>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, user-scalable=no" />
<link rel="stylesheet" href="assets/css/main.css" />
<!--[if lte IE 9]><link rel="stylesheet" href="assets/css/ie9.css" /><![endif]-->
<noscript><link rel="stylesheet" href="assets/css/noscript.css" /></noscript>
</head>
<body>
<!-- Wrapper -->
<div id="wrapper">
<!-- Header -->
<header id="header">
<div class="logo">
<span class="icon fa-users"></span>
</div>
<div class="content">
<div class="inner">
<h1>6PAC</h1>
<p>Making Probably Approximately Correct Learning <br>
Active, Sequential, Structure-aware, Efficient, Ideal and Safe</p>
<p>Joint research team between <a href="https://www.cwi.nl">CWI</a> and <a href="https://www.inria.fr">Inria</a></p>
</div>
</div>
<nav>
<ul>
<li><a href="#intro">Presentation</a></li>
<li><a href="#team">Team</a></li>
<li><a href="#events">Events</a></li>
<li><a href="#papers">Papers</a></li>
<li><a href="#contact">Contact</a></li>
<!--<li><a href="#elements">Elements</a></li>-->
</ul>
</nav>
</header>
<!-- Main -->
<div id="main">
<!-- Intro -->
<article id="intro">
<h2 class="major">Intro</h2>
<span class="image main"><img src="images/pic01.jpg" alt="" /></span>
<p>
See the <a href="https://www.inria.fr/en/associate-team/6pac">official announcement</a> on Inria's website.
</p>
<p style="text-align:justify">
This project roots in statistical learning theory, which can be viewed as the theoretical foundations of machine learning. The most common framework is a setup in which one is given n training examples, and the goal is to build a predictor that would be efficient on new (similar) data. This efficiency should be supported by PAC (Probably Approximately Correct) guarantees, e.g. upper bounds on the excess risk of a predictor that hold with high probability. Such guarantees however often hold under stringent assumptions which are typically never met in real-life application, e.g., independent, identically distributed data. More realistic modelling of data has triggered many research efforts in several directions: first, accommodating possible data (e.g., dependent, heavy-tailed), and second, in the direction of sequential learning, in which the predictor can be built on the fly, while new data is gathered. We believe that an ever more realistic paradigm is active learning, a setup in which the learner actively requests data (possibly facing constraints, such as storage, velocity, cost, etc.) and adapts its queries to optimize its performance. The 3-years objective of 6PAC (where 6 stands for Sequential, Active, Efficient, Structured, Ideal, Safe - the six research directions we intend to contribute to) is to pave the way to new PAC generalization and sample-complexity upper and lower bounds beyond batch learning. Our ambition is to contribute to several learning setups, ranging from sequential learning (where data streams are collected) to adaptive and active learning (where data streams are requested by the learning algorithm).
</p>
<!-- <p>Aenean ornare velit lacus, ac varius enim ullamcorper eu. Proin aliquam facilisis ante interdum congue. Integer mollis, nisl amet convallis, porttitor magna ullamcorper, amet egestas mauris. Ut magna finibus nisi nec lacinia. Nam maximus erat id euismod egestas. By the way, check out my <a href="#work">awesome work</a>.</p>
<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis dapibus rutrum facilisis. Class aptent taciti sociosqu ad litora torquent per conubia nostra, per inceptos himenaeos. Etiam tristique libero eu nibh porttitor fermentum. Nullam venenatis erat id vehicula viverra. Nunc ultrices eros ut ultricies condimentum. Mauris risus lacus, blandit sit amet venenatis non, bibendum vitae dolor. Nunc lorem mauris, fringilla in aliquam at, euismod in lectus. Pellentesque habitant morbi tristique senectus et netus et malesuada fames ac turpis egestas. In non lorem sit amet elit placerat maximus. Pellentesque aliquam maximus risus, vel sed vehicula.</p> -->
</article>
<!-- Team -->
<article id="team">
<h2 class="major">Team</h2>
<span class="image main"><img src="images/crew.jpg" alt="" /></span>
6PAC is a joint research team between <a href="https://www.cwi.nl">CWI</a> and <a href="https://www.inria.fr">Inria</a> and includes 4 researchers.
<ul>
<li>
<a href="https://safestatistics.com">Peter Grünwald</a>, co-PI
</li>
<li>
<a href="https://bguedj.github.io">Benjamin Guedj</a>, co-PI
</li>
<li>
<a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>
</li>
<li>
<a href="http://wouterkoolen.info">Wouter M. Koolen</a>
</li>
</ul>
<!-- <p>Adipiscing magna sed dolor elit. Praesent eleifend dignissim arcu, at eleifend sapien imperdiet ac. Aliquam erat volutpat. Praesent urna nisi, fringila lorem et vehicula lacinia quam. Integer sollicitudin mauris nec lorem luctus ultrices.</p>
<p>Nullam et orci eu lorem consequat tincidunt vivamus et sagittis libero. Mauris aliquet magna magna sed nunc rhoncus pharetra. Pellentesque condimentum sem. In efficitur ligula tate urna. Maecenas laoreet massa vel lacinia pellentesque lorem ipsum dolor. Nullam et orci eu lorem consequat tincidunt. Vivamus et sagittis libero. Mauris aliquet magna magna sed nunc rhoncus amet feugiat tempus.</p> -->
</article>
<!-- events -->
<article id="events">
<h2 class="major">Events</h2>
<ol>
<li>
<a href="http://wouterkoolen.info/">Wouter Koolen</a> visited Inria Lille - Nord Europe for about two weeks in May 2018, to work with <a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>. Their work has been accepted to <a href="https://nips.cc/Conferences/2018/Schedule?showEvent=11613">NeurIPS 2018</a>.
</li>
<li>
<a href="https://bguedj.github.io">Benjamin Guedj</a> and <a href="http://wouterkoolen.info/">Wouter Koolen</a> gave talks at the <a href="https://project.inria.fr/inriacwi/workshop-2018/">2nd CWI-Inria workshop</a>, at the Inria Paris research center on September 25--26, 2018.<br>
<a href="https://project.inria.fr/inriacwi/files/2018/11/BenjaminGuedj-6PAC-talk-sept2018.pdf">Benjamin's talk</a>: "A quasi-Bayesian perspective to Machine Learning"
<br>
<a href="https://project.inria.fr/inriacwi/files/2018/11/WouterKoolen-6PAC-talk-sep2018.pdf">Wouter's talk</a>: "Sequential Test for the Lowest Mean: From Thompson to Murphy Sampling"
</li>
<li>
<a href="http://wouterkoolen.info/">Wouter Koolen</a> visited Inria Lille - Nord Europe for about a week in September-October 2018, to work with <a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>.
</li>
<li>
<a href="https://safestatistics.com">Peter Grünwald</a> gave four 2-hours lectures at Inria Lille - Nord Europe in November 2018. Attendance of about 50 people.
</li>
<li>
<a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a> visited CWI for about a week in October 2018 to work with <a href="http://wouterkoolen.info/">Wouter Koolen</a>.
</li>
<li>
<a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a> visited CWI for about a week in February 2019 to work with <a href="http://wouterkoolen.info/">Wouter Koolen</a>.
</li>
<li>
<a href="https://homepages.cwi.nl/~heide/">Rianne de Heide</a> (PhD student, CWI, advised by <a href="https://safestatistics.com">Peter Grünwald</a> and <a href="http://wouterkoolen.info/">Wouter Koolen</a>) visited Inria Lille - Nord Europe from April 2019 to August 2019, to work with <a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>.
</li>
</ol>
<span class="image main"><img src="images/cwi-inria-workshop-2018.jpg" alt="" /></span>
<span class="image main"><img src="images/workshop2018wouter.jpg" alt="" /></span>
<span class="image main"><img src="images/workshop2018ben.jpg" alt="" /></span>
<span class="image main"><img src="images/peter_lecture.jpg" alt="" /></span>
<span class="image main"><img src="images/cheers.jpg" alt="" /></span>
<!-- <p>TBA</p> -->
<!-- <p>Lorem ipsum dolor sit amet, consectetur et adipiscing elit. Praesent eleifend dignissim arcu, at eleifend sapien imperdiet ac. Aliquam erat volutpat. Praesent urna nisi, fringila lorem et vehicula lacinia quam. Integer sollicitudin mauris nec lorem luctus ultrices. Aliquam libero et malesuada fames ac ante ipsum primis in faucibus. Cras viverra ligula sit amet ex mollis mattis lorem ipsum dolor sit amet.</p> -->
</article>
<!-- papers -->
<article id="papers">
<h2 class="major">Papers</h2>
<!-- <span class="image main"><img src="images/pic03.jpg" alt="" /></span> -->
<p>This page contains the list of research papers related to the 6PAC project.</p>
<ol>
<li>
<a href="https://arxiv.org/pdf/1806.00973.pdf">Sequential test for the lowest mean: From Thompson to Murphy sampling</a> <br>
<a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>, <a href="http://wouterkoolen.info">Wouter M. Koolen</a>, <a href="https://www.math.univ-toulouse.fr/~agarivie/?q=node/34">Aurélien Garivier</a><br>
Advances in Neural Information Processing Systems (NeurIPS), 2018
</li>
<li>
<a href="https://arxiv.org/abs/1811.11419">Mixture Martingales Revisited with Applications to Sequential Tests and Confidence Intervals</a> <br>
<a href="http://chercheurs.lille.inria.fr/ekaufman/">Emilie Kaufmann</a>, <a href="http://wouterkoolen.info">Wouter M. Koolen</a> <br>
Preprint
</li>
<li>
<a href="https://arxiv.org/abs/1906.10431">Non-Asymptotic Pure Exploration by Solving Games</a> <br>
Rémy Degenne, <a href="http://wouterkoolen.info">Wouter M. Koolen</a>, Pierre Ménard<br>
Advances in Neural Information Processing Systems (NeurIPS), 2019
</li>
<li>
<a href="https://arxiv.org/abs/1902.03475">Pure Exploration with Multiple Correct Answers</a> <br>
Rémy Degenne, <a href="http://wouterkoolen.info">Wouter M. Koolen</a><br>
Advances in Neural Information Processing Systems (NeurIPS), 2019
</li>
<li>
<a href="https://arxiv.org/abs/1905.13367">PAC-Bayes Un-Expected Bernstein Inequality</a> <br>
Zakaria Mhammedi, <a href="https://safestatistics.com">Peter Grünwald</a>, <a href="https://bguedj.github.io">Benjamin Guedj</a> <br>
Advances in Neural Information Processing Systems (NeurIPS), 2019
</li>
<li>
<a href="https://arxiv.org/abs/1905.10259">Dichotomize and Generalize: PAC-Bayesian Binary Activated Deep Neural Networks</a> <br>
Gaël Letarte, Pascal Germain, <a href="https://bguedj.github.io">Benjamin Guedj</a>, François Laviolette <br>
Advances in Neural Information Processing Systems (NeurIPS), 2019
</li>
</ol>
<!-- <p>Lorem ipsum dolor sit amet, consectetur et adipiscing elit. Praesent eleifend dignissim arcu, at eleifend sapien imperdiet ac. Aliquam erat volutpat. Praesent urna nisi, fringila lorem et vehicula lacinia quam. Integer sollicitudin mauris nec lorem luctus ultrices. Aliquam libero et malesuada fames ac ante ipsum primis in faucibus. Cras viverra ligula sit amet ex mollis mattis lorem ipsum dolor sit amet.</p> -->
</article>
<!-- Contact -->
<article id="contact">
<h2 class="major">Contact</h2>
<p>
<a href="https://bguedj.github.io">Link to Benjamin Guedj's webpage</a>.<br>
<a href="https://safestatistics.com">Link to Peter Grünwald's webpage</a>.
</p>
<!-- <form method="post" action="#">
<div class="field half first">
<label for="name">Name</label>
<input type="text" name="name" id="name" />
</div>
<div class="field half">
<label for="email">Email</label>
<input type="text" name="email" id="email" />
</div>
<div class="field">
<label for="message">Message</label>
<textarea name="message" id="message" rows="4"></textarea>
</div>
<ul class="actions">
<li><input type="submit" value="Send Message" class="special" /></li>
<li><input type="reset" value="Reset" /></li>
</ul>
</form> -->
<!-- <ul class="icons">
<li><a href="#" class="icon fa-twitter"><span class="label">Twitter</span></a></li>
<li><a href="#" class="icon fa-facebook"><span class="label">Facebook</span></a></li>
<li><a href="#" class="icon fa-instagram"><span class="label">Instagram</span></a></li>
<li><a href="#" class="icon fa-github"><span class="label">GitHub</span></a></li>
</ul> -->
</article>
<!-- Elements -->
<!-- <article id="elements">
<h2 class="major">Elements</h2>
<section>
<h3 class="major">Text</h3>
<p>This is <b>bold</b> and this is <strong>strong</strong>. This is <i>italic</i> and this is <em>emphasized</em>.
This is <sup>superscript</sup> text and this is <sub>subscript</sub> text.
This is <u>underlined</u> and this is code: <code>for (;;) { ... }</code>. Finally, <a href="#">this is a link</a>.</p>
<hr />
<h2>Heading Level 2</h2>
<h3>Heading Level 3</h3>
<h4>Heading Level 4</h4>
<h5>Heading Level 5</h5>
<h6>Heading Level 6</h6>
<hr />
<h4>Blockquote</h4>
<blockquote>Fringilla nisl. Donec accumsan interdum nisi, quis tincidunt felis sagittis eget tempus euismod. Vestibulum ante ipsum primis in faucibus vestibulum. Blandit adipiscing eu felis iaculis volutpat ac adipiscing accumsan faucibus. Vestibulum ante ipsum primis in faucibus lorem ipsum dolor sit amet nullam adipiscing eu felis.</blockquote>
<h4>Preformatted</h4>
<pre><code>i = 0;
while (!deck.isInOrder()) {
print 'Iteration ' + i;
deck.shuffle();
i++;
}
print 'It took ' + i + ' iterations to sort the deck.';</code></pre>
</section>
<section>
<h3 class="major">Lists</h3>
<h4>Unordered</h4>
<ul>
<li>Dolor pulvinar etiam.</li>
<li>Sagittis adipiscing.</li>
<li>Felis enim feugiat.</li>
</ul>
<h4>Alternate</h4>
<ul class="alt">
<li>Dolor pulvinar etiam.</li>
<li>Sagittis adipiscing.</li>
<li>Felis enim feugiat.</li>
</ul>
<h4>Ordered</h4>
<ol>
<li>Dolor pulvinar etiam.</li>
<li>Etiam vel felis viverra.</li>
<li>Felis enim feugiat.</li>
<li>Dolor pulvinar etiam.</li>
<li>Etiam vel felis lorem.</li>
<li>Felis enim et feugiat.</li>
</ol>
<h4>Icons</h4>
<ul class="icons">
<li><a href="#" class="icon fa-twitter"><span class="label">Twitter</span></a></li>
<li><a href="#" class="icon fa-facebook"><span class="label">Facebook</span></a></li>
<li><a href="#" class="icon fa-instagram"><span class="label">Instagram</span></a></li>
<li><a href="#" class="icon fa-github"><span class="label">Github</span></a></li>
</ul>
<h4>Actions</h4>
<ul class="actions">
<li><a href="#" class="button special">Default</a></li>
<li><a href="#" class="button">Default</a></li>
</ul>
<ul class="actions vertical">
<li><a href="#" class="button special">Default</a></li>
<li><a href="#" class="button">Default</a></li>
</ul>
</section>
<section>
<h3 class="major">Table</h3>
<h4>Default</h4>
<div class="table-wrapper">
<table>
<thead>
<tr>
<th>Name</th>
<th>Description</th>
<th>Price</th>
</tr>
</thead>
<tbody>
<tr>
<td>Item One</td>
<td>Ante turpis integer aliquet porttitor.</td>
<td>29.99</td>
</tr>
<tr>
<td>Item Two</td>
<td>Vis ac commodo adipiscing arcu aliquet.</td>
<td>19.99</td>
</tr>
<tr>
<td>Item Three</td>
<td> Morbi faucibus arcu accumsan lorem.</td>
<td>29.99</td>
</tr>
<tr>
<td>Item Four</td>
<td>Vitae integer tempus condimentum.</td>
<td>19.99</td>
</tr>
<tr>
<td>Item Five</td>
<td>Ante turpis integer aliquet porttitor.</td>
<td>29.99</td>
</tr>
</tbody>
<tfoot>
<tr>
<td colspan="2"></td>
<td>100.00</td>
</tr>
</tfoot>
</table>
</div>
<h4>Alternate</h4>
<div class="table-wrapper">
<table class="alt">
<thead>
<tr>
<th>Name</th>
<th>Description</th>
<th>Price</th>
</tr>
</thead>
<tbody>
<tr>
<td>Item One</td>
<td>Ante turpis integer aliquet porttitor.</td>
<td>29.99</td>
</tr>
<tr>
<td>Item Two</td>
<td>Vis ac commodo adipiscing arcu aliquet.</td>
<td>19.99</td>
</tr>
<tr>
<td>Item Three</td>
<td> Morbi faucibus arcu accumsan lorem.</td>
<td>29.99</td>
</tr>
<tr>
<td>Item Four</td>
<td>Vitae integer tempus condimentum.</td>
<td>19.99</td>
</tr>
<tr>
<td>Item Five</td>
<td>Ante turpis integer aliquet porttitor.</td>
<td>29.99</td>
</tr>
</tbody>
<tfoot>
<tr>
<td colspan="2"></td>
<td>100.00</td>
</tr>
</tfoot>
</table>
</div>
</section>
<section>
<h3 class="major">Buttons</h3>
<ul class="actions">
<li><a href="#" class="button special">Special</a></li>
<li><a href="#" class="button">Default</a></li>
</ul>
<ul class="actions">
<li><a href="#" class="button">Default</a></li>
<li><a href="#" class="button small">Small</a></li>
</ul>
<ul class="actions">
<li><a href="#" class="button special icon fa-download">Icon</a></li>
<li><a href="#" class="button icon fa-download">Icon</a></li>
</ul>
<ul class="actions">
<li><span class="button special disabled">Disabled</span></li>
<li><span class="button disabled">Disabled</span></li>
</ul>
</section>
<section>
<h3 class="major">Form</h3>
<form method="post" action="#">
<div class="field half first">
<label for="demo-name">Name</label>
<input type="text" name="demo-name" id="demo-name" value="" placeholder="Jane Doe" />
</div>
<div class="field half">
<label for="demo-email">Email</label>
<input type="email" name="demo-email" id="demo-email" value="" placeholder="jane@untitled.tld" />
</div>
<div class="field">
<label for="demo-category">Category</label>
<div class="select-wrapper">
<select name="demo-category" id="demo-category">
<option value="">-</option>
<option value="1">Manufacturing</option>
<option value="1">Shipping</option>
<option value="1">Administration</option>
<option value="1">Human Resources</option>
</select>
</div>
</div>
<div class="field half first">
<input type="radio" id="demo-priority-low" name="demo-priority" checked>
<label for="demo-priority-low">Low</label>
</div>
<div class="field half">
<input type="radio" id="demo-priority-high" name="demo-priority">
<label for="demo-priority-high">High</label>
</div>
<div class="field half first">
<input type="checkbox" id="demo-copy" name="demo-copy">
<label for="demo-copy">Email me a copy</label>
</div>
<div class="field half">
<input type="checkbox" id="demo-human" name="demo-human" checked>
<label for="demo-human">Not a robot</label>
</div>
<div class="field">
<label for="demo-message">Message</label>
<textarea name="demo-message" id="demo-message" placeholder="Enter your message" rows="6"></textarea>
</div>
<ul class="actions">
<li><input type="submit" value="Send Message" class="special" /></li>
<li><input type="reset" value="Reset" /></li>
</ul>
</form>
</section>
</article> -->
</div>
<!-- Footer -->
<footer id="footer">
<p class="copyright">© <a href="https://bguedj.github.io">Benjamin Guedj</a>. Design: <a href="https://html5up.net">HTML5 UP</a>.</p>
</footer>
</div>
<!-- BG -->
<div id="bg"></div>
<!-- Scripts -->
<script src="assets/js/jquery.min.js"></script>
<script src="assets/js/skel.min.js"></script>
<script src="assets/js/util.js"></script>
<script src="assets/js/main.js"></script>
<!-- Google Analytics -->
<script>
(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
})(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
ga('create', 'UA-81430854-1', 'auto');
ga('send', 'pageview');
</script>
</body>
</html>