forked from mmistakes/so-simple-theme
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html~
336 lines (254 loc) · 17.4 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
---
title: ""
date:
layout: mainpage
categories:
tags:
- home
published: true
comments:
---
<p>
Since November 2019, I'm professor at <a href="https://www.espci.fr">ESPCI (École Supérieure de Physique et de Chimie Industrielles de la Ville de Paris)</a> and
</p>
<ul class="org-ul">
<li>my research affiliation is in the <a href="https://www.lamsade.dauphine.fr/wp/miles/">Miles project</a> of the <a href="https://www.lamsade.dauphine.fr/">LAMSADE (Laboratoire d'Analyse et de Modélisation de Systèmes pour l’Aide à la Décision)</a> in <a href="https://www.dauphine.psl.eu/">Paris-Dauphine University</a> and <a href="https://www.psl.eu/">PSL (Paris Sciences & Lettres)</a>.</li>
<li>Head of the <a href="https://psl.eu/en/programmes-gradues/programme-data">DATA program of PSL</a>.</li>
<li>Co-head of the <a href="https://www.lamsade.dauphine.fr/en/research/groups/data-science.html">Data Science group of the LAMSADE</a> with <a href="https://www.lamsade.dauphine.fr/~elhaddad/">Joyce El Haddad</a>.</li>
</ul>
<div id="outline-container-org6000ff2" class="outline-2">
<h2 id="org6000ff2">News</h2>
<div class="outline-text-2" id="text-org6000ff2">
<ul class="org-ul">
<li><a href="https://allauzen.github.io/research/positions">Research positions</a>: For September 2024, two postions on Frugality in machine learning and speech processing
<ul class="org-ul">
<li>1 PhD position</li>
<li>1 postdoc position</li>
</ul></li>
</ul>
<ul class="org-ul">
<li>LeBenchmark 2.0 is available: more data, two additional tasks. <a href="https://arxiv.org/abs/2309.05472">Check the preprint</a>.</li>
<li>With the same team as the ICML 2023 paper (see below), we published a pre-print paper on <a href="https://arxiv.org/abs/2309.16883">Lipschitz-Variance-Margin Tradeoff for Enhanced Randomized Smoothing</a>.</li>
<li>Our paper "<a href="https://arxiv.org/abs/2305.16173">Efficient Bound of Lipschitz Constant for Convolutional Layers by Gram Iteration</a>" is at ICML 2023 ! Congratulations to my colleagues: Blaise Delattre, Quentin Barthélemy and Alexandre Araujo.</li>
<li>Our paper on <a href="https://arxiv.org/pdf/2112.08458.pdf">Curriculum learning for data-driven modeling of dynamical systems</a> is now published in the European Physical Journal.</li>
<li>With Alex Araujo, Aaron Havens, Blaise Delattre, and Bin Hu, we have a paper accepted at ICLR 2023 (in the top 25%): <a href="https://openreview.net/forum?id=k71IGLC8cfc">A Unified Algebraic Perspective on Lipschitz Neural Networks</a></li>
<li>A very nice web-page on <a href="https://soda-inria.github.io/ken_embeddings/">Relational Data Embeddings by Alexis Cvetkov</a>.</li>
</ul>
<a href="https://soda-inria.github.io/ken_embeddings/">
<img src="/assets/figs/entity_types_with_names.png" alt="Entity mapping" style="width:300px; margin:0px auto;display:block"/>
</a>
<ul class="org-ul">
<li>With Alexis Cvetkov and Gaël Varoquaux from the DirtyData team, we
have a paper on <a href="https://hal.archives-ouvertes.fr/hal-03647434/file/final.pdf">Analytics on Non-Normalized Data Sources: More
Learning, Rather Than More Cleaning</a>.</li>
</ul>
<a href="https://hal.archives-ouvertes.fr/hal-03647434/file/final.pdf">
<img src="/assets/figs/entity.png" alt="Entity mapping" style="width:300px; margin:0px auto;display:block"/>
</a>
<ul class="org-ul">
<li>The E-SSL ANR Project is accepted and will start soon. E-SSL stands
for <b><b>"Efficient Self-Supervised Learning for Inclusive and
Innovative Speech Technologies"</b></b>. Bravo to Titouan Parcollet, the
LIA and the LIG ! <b><b>A funded PhD position on "Fair and Inclusive
Self-Supervised Learning for Speech"</b></b> will be co-supervised between
Paris (LAMSADE) and Grenoble (LIG).</li>
<li>With Laurent Meunier, Blaise Delattre et Alex Araujo, we have a paper at ICML 2022: <a href="https://arxiv.org/abs/2110.12690">A Dynamical System Perspective for Lipschitz Neural Networks</a>.</li>
<li>with the incredible team of LeBenchmark (from Grenoble and Avignon), we have a paper at <b>NeurIPS 2021</b> in the <i>datasets and benchmarks track</i> ! Visit <a href="https://github.com/LeBenchmark/github">our repos</a>.</li>
</ul>
<div id="table-of-contents" role="doc-toc">
<h2>Table of Contents</h2>
<div id="text-table-of-contents" role="doc-toc">
<ul>
<li><a href="#org6000ff2">News</a></li>
<li><a href="#org0befa5c">Publication</a></li>
<li><a href="#orgf8d1728">Main research interest</a></li>
<li><a href="#org1e50265">Older news</a></li>
<li><a href="#org3715ab8">Talks on deep-learning for NLP</a></li>
<li><a href="#org58727e3">Contact</a></li>
</ul>
</div>
</div>
</div>
</div>
<div id="outline-container-org0befa5c" class="outline-2">
<h2 id="org0befa5c">Publication</h2>
<div class="outline-text-2" id="text-org0befa5c">
<p>
See <a href="http://scholar.google.com/citations?hl=en&user=B2-gXkkAAAAJ">my Google Scholar page</a>
</p>
</div>
</div>
<div id="outline-container-orgf8d1728" class="outline-2">
<h2 id="orgf8d1728">Main research interest</h2>
<div class="outline-text-2" id="text-orgf8d1728">
<p>
Professor at ESPCI and Researcher in MILES Team of the LAMSADE, my research topics are Natural Language Processing and Deep learning for physical data
</p>
</div>
</div>
<div id="outline-container-org1e50265" class="outline-2">
<h2 id="org1e50265">Older news</h2>
<div class="outline-text-2" id="text-org1e50265">
<ul class="org-ul">
<li>With <a href="https://smontariol.github.io/">Syrielle Montariol</a>, we have a paper at ACL 2021: <i>Measure and Evaluation of Semantic Divergence across Two Languages</i>. The camera ready version will be soon available.</li>
<li><a href="https://gdr-tal.ls2n.fr/etal-2021/">The summer school ETAL</a> (from the GDR TAL of CNRS) will take place in Lannion (14 to 18 of June 2021). I'm very happy to teach two courses. Check out the program.</li>
<li><a href="https://arxiv.org/abs/2104.11462">LeBenchmark</a> is out: A Reproducible Framework for Assessing Self-Supervised Representation Learning from Speech. Comming soon at Interspeech 2021.</li>
<li><b><b>FlauBERT</b></b> and <b><b>Flue</b></b> are now available: <a href="https://github.com/getalp/Flaubert">check the github repository</a> or read <a href="https://arxiv.org/abs/1912.05372">our paper</a> on arxiv.
<ul class="org-ul">
<li><b><b>FLauBERT</b></b> is a a French BERT trained on a very large and heterogeneous French corpus</li>
<li><b><b>Flue</b></b> is an evaluation setup for French NLP systems similar to the popular GLUE benchmark</li>
</ul></li>
</ul>
<ul class="org-ul">
<li>Read our paper on <a href="https://arxiv.org/abs/1906.07672">Control of chaotic systems by Deep Reinforcement Learning</a>. Comming soon, the journal version accepted in The Royal Society’s physical sciences research journal. The journal page is <a href="https://royalsocietypublishing.org/doi/10.1098/rspa.2019.0351">this one</a>.</li>
</ul>
<a href="https://arxiv.org/abs/1906.07672">
<img src="/assets/figs/ks_h.png" alt="Kuramoto-Sivashinski" style="width:400px; margin:0px auto;display:block"/>
</a>
<ul class="org-ul">
<li><b><b><a href="https://allauzen.github.io/research/positions">Research positions are available for 2021</a>: 2 PhD positions</b></b>. Coming soon a postdoc position (closed for PhD)</li>
</ul>
<ul class="org-ul">
<li>With Aina Gari Soler and <a href="https://perso.limsi.fr/marianna/">Marianna Apidianaki</a>, we have a paper at
<i>*Sem 19</i> on <a href="https://www.aclweb.org/anthology/S19-1002">Word Usage Similarity Estimation with Sentence
Representations and Automatic Substitutes</a>.</li>
</ul>
<a href="https://www.aclweb.org/anthology/S19-1002.pdf">
<img src="/assets/figs/word_usage.png" alt="Word usage" style="width:300px; margin:0px auto;display:block"/>
</a>
<ul class="org-ul">
<li>See also, the paper in
the Semantic Deep Learning (SemDeep-5) workshop at IJCAI:
<a href="http://www.dfki.de/~declerck/semdeep-5/papers/wic_SemDeep-5_paper_4.pdf">LIMSI-MULTISEM at the IJCAI SemDeep-5 WiC Challenge: Context
Representations for Word Usage Similarity Estimation</a>.</li>
</ul>
<ul class="org-ul">
<li>Syrielle Montariol's paper on <a href="https://arxiv.org/abs/1909.01863">Empirigcal Study of Diachronic Word Embeddings for Scarce Data</a> is available.</li>
</ul>
<ul class="org-ul">
<li><a href="http://atala.org/content/apprentissage-profond-pour-le-traitement-automatique-des-langues">The special issue on Deep Learning for NLP</a> of the TAL journal is now
online, with 3 nice papers.</li>
<li>November the 23th, a talk at <a href="https://www.sciencesmaths-paris.fr/fr/horizon-maths-2018-intelligence-artificielle-957.htm">Horizon Maths 2018 : Intelligence
Artificielle</a>, organized by the FSMP.</li>
<li>With Matthieu Labeau : <a href="https://aclanthology.coli.uni-saarland.de/papers/C18-1261/c18-1261">Learning with Noise-Contrastive Estimation:
Easing training by learning to scale</a>, in COLING 2018. The paper was
one of the <a href="https://coling2018.org/coling-2018-best-papers/">"Area Chair Favorites"</a>. Meet Matthieu in Santa-Fe.</li>
<li>Invited Talk at the day "NLP and AI" of the French conference on AI
(PFIA) in Nancy (06/07/2018). My talk was about <i>Language models: large
vocabulary challenges</i>.</li>
<li>With Hinrich Schütze and Sophie Rosset, we are editors of a
Special issue on <b>Deep-Learning for NLP</b> of the TAL journal. Submit
a paper ! See <a href="https://tal-59-2.sciencesconf.org/">the official page for more detail</a>.</li>
<li>Our paper <a href="https://arxiv.org/abs/1603.05962">Document Neural Autoregressive Distribution Estimation</a>
with Stanislas Lauly, Yin Zheng and Hugo Larochelle has been
published in <b>JMLR</b>.</li>
<li>With Matthieu Labeau: <b>Best paper award</b> at <a href="https://sites.google.com/view/sclem2017/schedule"> SCLeM 2017</a> with our paper <i>Character and Subword-Based Word Representation for Neural Language Modeling Prediction</i>. If you attend EMNLP workshops, see the talk and the poster, September 7, 2017 in Copenhagen.</li>
<li>Opening talk at <a href="http://seminaire-dga.gforge.inria.fr/2016/DeuxiemeJourneedefistechnologiquesdelacybersecurite_fr.html">the DGA/IRISA seminar on Artificial Intelligence and Security</a>, 25/04/17. The slides are available <a href="https://perso.limsi.fr/Individu/doc/aa_rennes17.pdf">here</a>.</li>
<li>Our paper in Machine translation Journal on <b>A comparison of discriminative training criteria for continuous space translation models</b> is now <a href="http://link.springer.com/journal/10590">online</a></li>
</ul>
<ul class="org-ul">
<li>Matthieu Labeau presents in EACL our paper : <a href="http://www.aclweb.org/anthology/E/E17/E17-2003.pdf">An experimental analysis of Noise-Contrastive Estimation: the noise distribution matters</a></li>
<li>Invited talk at <a href="https://indico.math.cnrs.fr/event/1799/">the Workshop "Statistics/Learning at Paris-Saclay”</a>, Institut des hautes
études scientifiques (IHES), 19/01/2017. The slides are available <a href="https://perso.limsi.fr/Individu/doc/v0_aa_ihess.pdf">here</a>.</li>
<li>Our (French) paper in the journal TAL is now online:
<a href="http://www.atala.org/Apprentissage-discriminant-de">Apprentissage discriminant de modèles neuronaux pour la traduction automatique</a>.
You can check the <a href="http://www.atala.org/-Numero-1-">whole issue</a>
which contains other great papers.</li>
<li>The special issue on <b>Deep Learning for Machine Translation</b> in the
journal <b>Computer Speech and Language</b> should be online soon.</li>
<li>LIMSI Papers at the conference on statistical machine translation,
aka <a href="http://statmt.org/wmt17/">WMT17</a>.</li>
<li>Check the website of the Digicosme working group on <b><a href="https://gt-deepnet.limsi.fr/">Deep Nets and learning representation</a></b> (in French).</li>
</ul>
<ul class="org-ul">
<li>3 papers at the first conference on statistical machine translation,
aka <a href="http://statmt.org/wmt16/">WMT16</a>.</li>
</ul>
<ul class="org-ul">
<li>Tutorial on Deep learning and NLP applications at the
<a href="http://www.irit.fr/cimi-machine-learning/node/9">Workshop on Learning with Structured Data and applications on Natural Language and Biology</a>: the slides are available in
<a href="https://perso.limsi.fr/allauzen/doc/aa_deep_nlp.pdf">pdf</a> (the file file is quite large, about 9MB).</li>
<li>Two Papers at EMNLP this fall 2015
<ul class="org-ul">
<li><a href="http://www.aclweb.org/anthology/D/D15/D15-1025.pdf">Non-lexical neural architecture for fine-grained POS Tagging</a></li>
<li><a href="http://www.aclweb.org/anthology/D/D15/D15-1121.pdf">A Discriminative Training Procedure for Continuous Translation Models</a></li>
</ul></li>
<li>Three Papers at WMT (see <a href="http://statmt.org/wmt15/papers.html">this website</a>), including one with Jan Niehues on ListNet-based MT
Rescoring.</li>
<li><b>Prix Ex-aequo du meilleur article à TALN 2015</b>
<a href="http://www.atala.org/TALN-RECITAL-2015-22eme-conference">Apprentissage discriminant des modèles continus de traduction</a></li>
<li>Check the website of the Digicosme working group on
<b><a href="https://gt-deepnet.limsi.fr/">Deep Nets and learning representation</a></b> (in French). We will have nice invited speakers for
the upcoming sessions:
<ul class="org-ul">
<li>April the 8th: Antoine Bordes (Facebook)</li>
<li>April the 16th: Florence D'Alché-Buc and Romain Brault
(Telecom-ParisTech), along with a discussion lead by Aurélien
Decelle (LRI) on "Approximate Message Passing with Restricted
Boltzmann Machine Priors"</li>
<li>May the 22nd: Edward Grefenstette (Google-Deepmind)</li>
<li>June the 18th: Stéphane Mallat (CMAP)</li>
</ul></li>
</ul>
<ul class="org-ul">
<li>Think about submiting to the 3rd edition of the workshop on
<b><a href="https://sites.google.com/site/cvscworkshop2015">Continuous Vector Space Models and their Compositionality (CVSC)</a></b>, with a nice set of
Keynote speakers:
<ul class="org-ul">
<li>Kyunghyun Cho (Montreal)</li>
<li>Stephen Clark (Cambridge)</li>
<li>Yoav Goldberg (Bar Ilan)</li>
<li>Ray Mooney (Texas)</li>
<li>Jason Weston (Facebook AI Research)</li>
</ul></li>
<li>December 2014: Two papers at IWSLT, one of them is on
<a href="https://perso.limsi.fr/allauzen/doc/dokhanh_iwslt14.pdf">Discriminative Adaptation of Continuous Space Translation Models</a>.</li>
<li>There will be a third edition of our workshop on <b>Continuous Vector
Space Models and their Compositionality (CVSC)</b>, colocated with
<a href="http://acl2015.org/">ACL 2015</a> in Beijing.</li>
<li>October 2014: <a href="http://amta2014.amtaweb.org/Proceedings.aspx">AMTA paper</a> <i>Combining Techniques from different NN-based Language Models
for Machine Translation</i> with <a href="http://scholar.google.com/citations?hl=fr&user=fO9cszYAAAAJ">Jan Nieheus</a> is now published</li>
<li>July the 3rd: Two papers in <a href="http://www.taln2014.org/site/programme/programme-detaille-sessions-orales/">TALN</a></li>
<li>June 2014, <a href="http://www.statmt.org/wmt14/papers.html">WMT14</a>: 3 papers !</li>
<li>June the 3rd, 2014: talk at <a href="http://www.aftal.fr/jadt2014/?page_id=13">l'atelier corpus multilingues</a>, JADT 2014 (in French)</li>
<li>April 27th 2014: <b><a href="https://sites.google.com/site/cvscworkshop2014/">Workshop on Continuous Vector Space Models and their Compositionality</a></b>, this is the second edition.</li>
<li>March 25th 2014: talk at the joint one day workshop on <a href="http://www.afia.asso.fr/tiki-index.php?page=Journ%C3%A9e+commune+AFIA+-+ATALA+2014">"Langue, apprentissage automatique et fouille de données"</a> (in French)</li>
<li>Habilitation à diriger des recherches
(<a href="http://fr.wikipedia.org/wiki/Habilitation_universitaire">see this page to know what does it mean)</a>:
<ul class="org-ul">
<li>January the 30th, 2014 at LIMSI</li>
<li>Title: Statistical models for machine translation</li>
<li>my <a href="http://perso.limsi.fr/Individu/allauzen/doc/aa_slides.pdf">slides</a></li>
<li>my <a href="http://perso.limsi.fr/Individu/allauzen/doc/hdr_allauzen.pdf">report</a></li>
</ul></li>
</ul>
</div>
</div>
<div id="outline-container-org3715ab8" class="outline-2">
<h2 id="org3715ab8">Talks on deep-learning for NLP</h2>
<div class="outline-text-2" id="text-org3715ab8">
<ul class="org-ul">
<li>Opening talk at <a href="http://seminaire-dga.gforge.inria.fr/2016/DeuxiemeJourneedefistechnologiquesdelacybersecurite_fr.html">the DGA/IRISA seminar on Artificial Intelligence and Security</a>, 25/04/17. The slides are available <a href="https://perso.limsi.fr/Individu/allauzen/doc/aa_rennes17.pdf">here</a>.</li>
<li>Invited talk at <a href="https://indico.math.cnrs.fr/event/1799/">the Workshop "Statistics/Learning at Paris-Saclay”</a>, Institut des hautes
études scientifiques (IHES), 19/01/2017. The slides are available <a href="https://perso.limsi.fr/Individu/allauzen/doc/v0_aa_ihess.pdf">here</a>.</li>
<li>Tutorial on Deep learning and NLP applications at the
<a href="http://www.irit.fr/cimi-machine-learning/node/9">Workshop on Learning with Structured Data and applications on Natural Language and Biology</a>: the slides are available in
<a href="https://perso.limsi.fr/Individu/allauzen/doc/aa_deep_nlp.pdf">pdf</a> (the file file is quite large, about 9MB).</li>
</ul>
</div>
</div>
<div id="outline-container-org58727e3" class="outline-2">
<h2 id="org58727e3">Contact</h2>
<div class="outline-text-2" id="text-org58727e3">
<p>
My postmail address (email prefered):
</p>
<p>
Alexandre Allauzen
Université Dauphine - PSL, Laboratoire LAMSADE,
Place du Maréchal de Lattre de Tassigny,
75 775 Paris Cedex 16,
France.
</p>
</div>
</div>