francoispierrepaty.github.io/index.html at master · francoispierrepaty/francoispierrepaty.github.io · 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
<!doctype html>
<html lang="en">
  <head>
    <!-- Global site tag (gtag.js) - Google Analytics -->
    <script async src="https://www.googletagmanager.com/gtag/js?id=UA-133810388-1"></script>
    <script>
      window.dataLayer = window.dataLayer || [];
      function gtag(){dataLayer.push(arguments);}
      gtag('js', new Date());
      gtag('config', 'UA-133810388-1');
    </script>

    <meta charset="utf-8">
    <meta name="viewport" content="width=device-width, initial-scale=1">
    <title>François-Pierre Paty</title>
    <meta name="description" content="François-Pierre Paty. Applied mathematics, machine learning, optimal transport, time series and operations research.">

    <link rel="preconnect" href="https://fonts.googleapis.com">
    <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
    <link href="https://fonts.googleapis.com/css2?family=Spectral:ital,wght@0,400;0,500;0,600;1,400&display=swap" rel="stylesheet">
    <link rel="stylesheet" href="https://use.fontawesome.com/releases/v5.8.1/css/all.css" integrity="sha384-50oBUHEmvpQ+1lW4y57PTFmhCaXp0ML5d60M1M7uH2+nqUivzIebhndOJK28anvf" crossorigin="anonymous">
    <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/academicons/1.9.4/css/academicons.min.css">
    <link rel="stylesheet" href="stylesheets/main.css">
    <script>
      window.MathJax = { tex: { inlineMath: [['\\(', '\\)']], displayMath: [['$$', '$$']] }, svg: { fontCache: 'global' } };
    </script>
    <script src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-svg.js" async></script>
  </head>

  <body data-page="home">
    <div class="wrap">
      <div class="layout">

        <aside class="sidebar"><!-- injected by javascripts/sidebar.js --></aside>

        <main>
          <section id="about">
            <h2>About me</h2>
            <p class="lead">
              Right now I am a visiting professor at <a href="https://ecpknfr.buaa.edu.cn">École Centrale de Pékin</a>, in China, where I teach analysis and probability in French to Chinese students.
            </p>
            <p>
              Before that I was the first employee and R&amp;D Lead at <a href="https://www.ida.eu">Ida</a>, a French AI startup that optimizes fresh food orders for retailers. I built its whole R&amp;D pipeline from scratch, and Ida was bought by <a href="https://www.relexsolutions.com">RELEX Solutions</a> in 2025.
            </p>
            <p>
              On the side I run <a href="https://www.hsklevel.com">HSK Level</a> and <a href="https://www.chinesegrammargems.com">Chinese Grammar Gems</a>, a platform and newsletter for learning Chinese that together reach more than 25,000 people around the world.
            </p>
            <p>
              I did my PhD in applied mathematics and machine learning at <a href="https://www.ip-paris.fr/en">Institut Polytechnique de Paris</a>, working on algorithmic optimal transport with <a href="http://marcocuturi.net">Marco Cuturi</a> as my advisor.
            </p>
          </section>

          <section id="education">
            <h2>Education</h2>
            <ul class="tight">
              <li>PhD from <a href="https://www.ip-paris.fr/en">Institut Polytechnique de Paris</a></li>
              <li>Engineering degree from <a href="https://www.polytechnique.edu/en/ingenieur-polytechnicien-program">Ecole Polytechnique</a></li>
              <li>Engineering degree from <a href="https://www.ensae.fr/en/">ENSAE Paris</a></li>
              <li>Master's degree in <a href="https://www.imo.universite-paris-saclay.fr/fr/etudiants/masters/mathematiques-et-applications/m2/mathematiques-de-laleatoire/">Statistics and Machine Learning</a> from <a href="https://www.math.u-psud.fr">Université Paris-Sud</a></li>
            </ul>
          </section>

          <section id="research-teaser">
            <h2>Research interests</h2>
            <p>
              My background is in <em>algorithmic optimal transport</em>, with applications to <em>machine learning</em> and <em>economics</em>, and more recently to <em>time series, forecasting</em> and <em>operations research</em>.
              <a href="research.html">Read more, and see publications &amp; talks →</a>
            </p>
          </section>

          <!-- ===== Featured interactive demo ===== -->
          <section id="featured-demo">
            <h2>Featured demo · Subspace Robust Wasserstein</h2>
            <p>
              A live taste of my research: <a href="http://proceedings.mlr.press/v97/paty19a.html">Subspace Robust Wasserstein Distances</a> (Paty &amp; Cuturi, ICML 2019). Optimal transport measures the effort to morph one point cloud, \(\mu\), into another, \(\nu\); this demo finds the single line (yellow) along which the two clouds are <em>hardest</em> to align. Drag either cloud to move it around.
            </p>
            <div class="demo-card">
              <div class="demo-canvas-wrap">
                <canvas id="srw-canvas" width="900" height="520"></canvas>
              </div>
              <div class="demo-controls">
                <button class="btn" id="srw-resample">Resample clouds</button>
                <button class="btn ghost" id="srw-toggle-plan">Show transport plan</button>
                <label>Spread of Y
                  <input type="range" id="srw-spread" min="0.04" max="0.34" step="0.01" value="0.18">
                </label>
                <label>Regularization&nbsp;ε
                  <input type="range" id="srw-eps" min="0.0005" max="0.015" step="0.0005" value="0.003">
                </label>
              </div>
              <div class="demo-legend">
                <span><span class="swatch" style="background:#2a6f7a"></span>first cloud (μ)</span>
                <span><span class="swatch" style="background:#a4452f"></span>second cloud (ν)</span>
                <span><span class="line-swatch" style="border-color:#caa84a"></span>worst-case line (k = 1)</span>
                <span><span class="line-swatch dashed" style="border-color:#6a6258"></span>PCA</span>
              </div>
              <div class="demo-readout">
                <span>robust distance S₁: <b id="srw-val">…</b></span>
                <span>ordinary Wasserstein W₂: <b id="srw-w2">…</b></span>
                <span>angle of the worst-case line: <b id="srw-dir">…</b></span>
              </div>
            </div>
            <p class="more-link">
              <a href="demos.html#srw">Full explanation and the other interactive demos →</a>
            </p>
          </section>

          <section id="news">
            <h2>News</h2>
            <ul class="news">
              <li><strong>December 2025:</strong> <a href="https://www.ida.eu">Ida</a> was acquired by <a href="https://www.relexsolutions.com">RELEX Solutions</a>.</li>
              <li><strong>September 2025:</strong> I joined <a href="https://ecpknfr.buaa.edu.cn">École Centrale de Pékin</a> as a visiting mathematics professor.</li>
              <li><strong>December 11, 2024:</strong> I gave a course on operations research at the <a href="https://ecoledesponts.fr/en">Ecole Nationale des Ponts et Chaussées</a>.</li>
              <li><strong>October 2, 2023:</strong> I joined <a href="https://www.ida.eu">Ida</a> as a research scientist.</li>
            </ul>
            <details>
              <summary>Older events</summary>
              <ul class="news">
                <li><strong>June 27, 2022:</strong> I gave a talk in the <a href="http://angkor.univ-mlv.fr/~vialard/project/newmongeproblems/">New Monge Problems</a> seminar at the Université Gustave Eiffel near Paris.</li>
                <li><strong>May 11, 2022:</strong> I gave a talk (slides in <a href="assets/slides/Mokaplan.pdf">.pdf</a>) at the <a href="https://team.inria.fr/mokaplan/mokameetings/">Mokaplan team seminar</a> at INRIA in Paris.</li>
                <li><strong>October 21–22, 2021:</strong> I participated in the <em>Paris workshop on optimal transport with applications to economics and statistics</em> at Sciences Po Paris.</li>
                <li><strong>June 29, 2021:</strong> I defended my PhD thesis (<a href='assets/slides/phd_defense.pdf'>slides in pdf</a> and the <a href='https://tel.archives-ouvertes.fr/tel-03316856/document'>manuscript</a>) at ENSAE Paris.</li>
                <li><strong>March 31, 2021:</strong> I gave a talk at the <a href="https://www.imo.universite-paris-saclay.fr/-des-doctorants-">EDMH PhD students seminar</a> (<a href="https://scalelite.lal.cloud.math.cnrs.fr/playback/presentation/2.0/playback.html?meetingId=07328e5495597cd77347ced584539447d62e46a3-1617195878320">video in French</a>).</li>
                <li><strong>March 4, 2021:</strong> I gave a talk at the <a href="https://www.math.u-bordeaux.fr/imb/seminaire-image-optimisation-et-probabilites">Image, Optimization and Probability seminar</a> at the <a href="http://image.math.u-bordeaux.fr">Institut de Mathématiques de Bordeaux</a>.</li>
                <li><strong>August 28, 2020:</strong> We received a <strong>Notable Paper Award</strong> at <a href="http://aistats.org/aistats2020/index.html">AISTATS 2020</a> for our paper <em>"<a href="http://proceedings.mlr.press/v108/paty20a.html">Regularity as Regularization: Smooth and Strongly Convex Brenier Potentials in Optimal Transport</a>"</em>.</li>
                <li><strong>August 26–28, 2020:</strong> I presented our paper <em>"<a href="http://proceedings.mlr.press/v108/paty20a.html">Regularity as Regularization: Smooth and Strongly Convex Brenier Potentials in Optimal Transport</a>"</em> at <a href="http://aistats.org/aistats2020/index.html">AISTATS 2020</a>.</li>
                <li><strong>July 12–18, 2020:</strong> I presented our paper <em>"<a href="https://arxiv.org/abs/2002.03967">Regularized Optimal Transport is Ground Cost Adversarial</a>"</em> at <a href="https://icml.cc">ICML 2020</a>.</li>
                <li><strong>January 17, 2020:</strong> I gave a talk (<a href="assets/slides/presentation_seminar_CEA.pdf">slides</a>) at the seminar day <a href="https://www.cosmostat.org/events/learning-in-astrophysics-2020">Learning meets Astrophysics</a> at CEA in Saclay, France.</li>
                <li><strong>December 08–14, 2019:</strong> I presented our paper <em>Regularity as Regularization: Smooth and Strongly Convex Brenier Potentials</em> at NeurIPS <a href="https://otml2019.github.io/">Optimal Transport and Machine Learning Workshop</a> in Vancouver, Canada.</li>
                <li><strong>November 20, 2019:</strong> I gave an introductory lecture on optimal transport at the seminar <a href="https://StatEcoML.github.io">Stat·Eco·ML</a> in Palaiseau, France.</li>
                <li><strong>November 05, 2019:</strong> I gave a talk (<a href="assets/slides/presentation_seminaire_palaisien.pdf">slides</a>) at <a href="https://palaisien.herokuapp.com/">Le Séminaire Palaisien</a> in Palaiseau, France.</li>
                <li><strong>August 26–31, 2019:</strong> I participated in, and gave a tutorial (<a href="https://github.com/mlss-skoltech/tutorials/tree/master/optimal_transport_tutorial">notebooks</a>, <a href="https://www.youtube.com/watch?v=lHrLAkYeHkk">video</a>) at <a href="https://mlss2019.skoltech.ru/">Machine Learning Summer School</a> in Moscow, Russian Federation.</li>
                <li><strong>July 07–19, 2019:</strong> I participated in, gave a talk (<a href="assets/slides/presentation_saint_flour.pdf">.pdf</a>) and presented a poster (<a href="assets/posters/poster_saint_flour.pdf">.pdf</a>), in <a href="http://recherche.math.univ-bpclermont.fr/stflour/">Saint-Flour Probability Summer School</a> in Saint-Flour, France.</li>
                <li><strong>June 24–28, 2019:</strong> I participated in the workshop <a href="http://web.math.unifi.it/users/depascal/Palazzone2019/">People in Optimal Transportation and Applications</a> in Cortona, Italy.</li>
                <li><strong>June 09–15, 2019:</strong> I presented (<a href="assets/posters/poster_ICML19.pdf">poster</a> and <a href="https://slideslive.com/38917398/general-ml">20-minute oral</a>) our paper <em>"<a href="http://proceedings.mlr.press/v97/paty19a.html">Subspace Robust Wasserstein Distances</a>"</em> at <a href="https://icml.cc/Conferences/2019">ICML 2019</a> in Long Beach, USA.</li>
                <li><strong>March 25–29, 2019:</strong> I presented a poster (<a href="assets/posters/poster_les_houches.pdf">.pdf</a>) at the workshop <a href="https://ljk.imag.fr/membres/Jerome.Malick/osl2019/index.html">Optimization and Statistical Learning</a> in Les Houches, France.</li>
              </ul>
            </details>
          </section>

          <section id="intro">
            <h2>Introduction au transport optimal (en français)</h2>
            <p>Voici une introduction au transport optimal extraite d'un <a href="https://scalelite.lal.cloud.math.cnrs.fr/playback/presentation/2.0/playback.html?meetingId=07328e5495597cd77347ced584539447d62e46a3-1617195878320">séminaire de vulgarisation</a> que j'ai donné à destination des doctorants de l'École doctorale Hadamard en mars 2021.</p>
            <div class="video">
              <iframe src="https://www.youtube.com/embed/lx8FTjw02RM" title="Introduction au transport optimal" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
            </div>
          </section>

        </main>
      </div>
    </div>
    <script src="javascripts/sidebar.js"></script>
    <script src="javascripts/srw.js"></script>
  </body>
</html>