-
Notifications
You must be signed in to change notification settings - Fork 0
/
paper_2.html
382 lines (275 loc) · 17.2 KB
/
paper_2.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
<!DOCTYPE html>
<html lang="en">
<head>
<!-- Required meta tags -->
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no" />
<!-- External Javascript libs_ext -->
<script src="https://cdn.jsdelivr.net/npm/d3@5/dist/d3.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/handlebars@4.7.3/dist/handlebars.min.js"
integrity="sha256-/PJBs6QWvXijOFIX04kZpLb6ZtSQckdOIavLWKKOgXU=" crossorigin="anonymous"></script>
<script src="https://code.jquery.com/jquery-3.6.0.min.js"
integrity="sha256-/xUj+3OJU5yExlq6GSYGSHk7tPXikynS7ogEvDej/m4=" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/popper.js@1.16.0/dist/umd/popper.min.js"
integrity="sha384-Q6E9RHvbIyZFJoft+2mJbHaEWldlvI9IOYy5n3zV9zzTtmI3UksdQRVvoxMfooAo"
crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/bootstrap@4.4.1/dist/js/bootstrap.min.js"
integrity="sha256-WqU1JavFxSAMcLP2WIOI+GB2zWmShMI82mTpLDcqFUg=" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/moment@2.24.0/min/moment.min.js"
integrity="sha256-4iQZ6BVL4qNKlQ27TExEhBN1HFPvAvAMbFavKKosSWQ=" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/moment-timezone@0.5.28/builds/moment-timezone-with-data.min.js"
integrity="sha256-IWYg4uIC8/erItNXYvLtyYHioRi2zT1TFva8qaAU/ww=" crossorigin="anonymous"></script>
<!-- Library libs_ext -->
<script src="static/js/libs_ext/typeahead.bundle.js"></script>
<script src="static/js/tagEditor/jquery.caret.min.js"></script>
<script src="static/js/tagEditor/jquery.tag-editor.min.js"></script>
<!-- Internal Libs -->
<script src="static/js/data/api.js"></script>
<!-- External CSS -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/bootstrap@4.3.1/dist/css/bootstrap.min.css"
integrity="sha256-YLGeXaapI0/5IgZopewRJcFXomhRMlYYjugPLSyNjTY=" crossorigin="anonymous">
<!-- External Fonts (no google for china) -->
<link href="static/css/Lato.css" rel="stylesheet" />
<link href="static/css/Exo.css" rel="stylesheet" />
<link href="static/css/Cuprum.css" rel="stylesheet" />
<link rel="stylesheet" href="static/css/main.css" />
<!-- <link rel="stylesheet" href="static/css/fa_regular.css"/>-->
<link rel="stylesheet" href="static/css/fa_solid.css" />
<link rel="stylesheet" href="static/css/lazy_load.css" />
<link rel="stylesheet" href="static/css/typeahead.css" />
<script type="text/javascript" src="https://cdn.jsdelivr.net/npm/daterangepicker/daterangepicker.min.js"></script>
<link rel="stylesheet" type="text/css" href="https://cdn.jsdelivr.net/npm/daterangepicker/daterangepicker.css" />
<title>Machine Learning Simulation: Learning Mesh-Based Simulation with Graph Networks</title>
<link rel="icon" type="image/png" href="favicon.ico">
<!-- Google Analytics -->
<!-- <script async src="https://www.googletagmanager.com/gtag/js?id=G-GT7LEVW06Z"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() { dataLayer.push(arguments); }
gtag('js', new Date());
gtag('config', 'G-GT7LEVW06Z');
</script> -->
<!-- head block ends here-->
<meta name="citation_title" content="Learning Mesh-Based Simulation with Graph Networks" />
<meta name="citation_author" content="Tobias Pfaff" />
<meta name="citation_author" content="Meire Fortunato" />
<meta name="citation_author" content="Alvaro Sanchez-Gonzalez" />
<meta name="citation_author" content="Peter W. Battaglia" />
<meta name="citation_abstract" content="Mesh-based simulations are central to modeling complex physical systems in many disciplines across science and engineering. Mesh representations support powerful numerical integration methods and their resolution can be adapted to strike favorable trade-offs between accuracy and efficiency. However, high-dimensional scientific simulations are very expensive to run, and solvers and parameters must often be tuned individually to each system studied. Here we introduce MeshGraphNets, a framework for learning mesh-based simulations using graph neural networks. Our model can be trained to pass messages on a mesh graph and to adapt the mesh discretization during forward simulation. Our results show it can accurately predict the dynamics of a wide range of physical systems, including aerodynamics, structural mechanics, and cloth. The model's adaptivity supports learning resolution-independent dynamics and can scale to more complex state spaces at test time. Our method is also highly efficient, running 1-2 orders of magnitude faster than the simulation on which it is trained. Our approach broadens the range of problems on which neural network simulators can operate and promises to improve the efficiency of complex, scientific modeling tasks." />
<meta name="citation_keywords" content="MeshGraphNets" />
<meta name="citation_keywords" content="Graph Neural Networks" />
<meta name="citation_keywords" content="forward simulation" />
<meta name="citation_keywords" content="GNNs" />
<meta name="citation_keywords" content="message passing" />
<meta name="citation_keywords" content="adaptive remeshing" />
<meta name="citation_keywords" content="encoder-processor-decoder" />
<meta name="citation_keywords" content="physical simulation" />
<meta name="citation_keywords" content="runtime efficiency" />
<meta name="citation_keywords" content="neural network simulators" />
<meta name="citation_pdf_url" content="https://arxiv.org/pdf/2010.03409.pdf" />
<script type="text/javascript" src="static/js/vis/vis-network.min.js"></script>
<style>
.red-hyper-link {
color: #ED1C24 !important;
}
</style>
</head>
<body>
<!-- NAV -->
<!-- add the below items to the "navigation_bar" list if those pages are needed; can't comment those out b/c jinja doesn't support comment
('index.html', 'Home'),
('papers.html', 'Advanced Search'),
('add_paper.html', 'Add Paper'),
('join_us.html', 'Join Us'),
('faq.html', 'FAQ')
-->
<nav class="navbar sticky-top navbar-expand-lg navbar-light" style="background-color: #EFECE5;" id="main-nav">
<div class="container">
<a class="navbar-brand" href="index.html">
<img class="logo" src="static/images/mls_logo_withfont.svg" style="width: 250px;"/>
</a>
<button class="navbar-toggler" type="button" data-toggle="collapse" data-target="#navbarNav"
aria-controls="navbarNav" aria-expanded="false" aria-label="Toggle navigation">
<span class="navbar-toggler-icon"></span>
</button>
<div class="collapse navbar-collapse text-right flex-grow-1" id="navbarNav">
<ul class="navbar-nav ml-auto">
<li class="nav-item ">
<a class="nav-link" href="index.html" style="font-family: 'Ubuntu', sans-serif;">Home</a>
</li>
<li class="nav-item ">
<a class="nav-link" href="add_paper.html" style="font-family: 'Ubuntu', sans-serif;">Add Paper</a>
</li>
</ul>
</div>
</div>
</nav>
<!-- User Overrides -->
<!-- top block ends here-->
<div class="container">
<!-- Tabs -->
<div class="tabs">
</div>
<!-- Content -->
<div class="content">
<!-- Title -->
<div class="pp-card m-3">
<div class="card-header">
<h2 class="card-title main-title text-center" style="color: black;">
Learning Mesh-Based Simulation with Graph Networks
</h2>
<h3 class="card-subtitle mb-2 text-muted text-center">
<a href="papers.html?author=Tobias Pfaff" target="_blank"
data-tippy-content="See all papers authored by Tobias Pfaff"
class="text-muted filterByAuthorLink">Tobias Pfaff</a>,
<a href="papers.html?author=Meire Fortunato" target="_blank"
data-tippy-content="See all papers authored by Meire Fortunato"
class="text-muted filterByAuthorLink">Meire Fortunato</a>,
<a href="papers.html?author=Alvaro Sanchez-Gonzalez" target="_blank"
data-tippy-content="See all papers authored by Alvaro Sanchez-Gonzalez"
class="text-muted filterByAuthorLink">Alvaro Sanchez-Gonzalez</a>,
<a href="papers.html?author=Peter W. Battaglia" target="_blank"
data-tippy-content="See all papers authored by Peter W. Battaglia"
class="text-muted filterByAuthorLink">Peter W. Battaglia</a>
</h3>
<h3 class="card-subtitle mb-2 text-muted text-center">
18/6/2021
</h3>
<div id="citation-code-wrapper" class="text-muted text-center"
>
</div>
<p class="text-center" style="margin-bottom: 0px;">
<span>Keywords:</span>
<a href="papers.html?keyword=MeshGraphNets" target="_blank"
data-tippy-content="See all papers with keyword 'MeshGraphNets'"
class="text-secondary text-decoration-none filterByKeywordLink">MeshGraphNets</a>,
<a href="papers.html?keyword=Graph Neural Networks" target="_blank"
data-tippy-content="See all papers with keyword 'Graph Neural Networks'"
class="text-secondary text-decoration-none filterByKeywordLink">Graph Neural Networks</a>,
<a href="papers.html?keyword=forward simulation" target="_blank"
data-tippy-content="See all papers with keyword 'forward simulation'"
class="text-secondary text-decoration-none filterByKeywordLink">forward simulation</a>,
<a href="papers.html?keyword=GNNs" target="_blank"
data-tippy-content="See all papers with keyword 'GNNs'"
class="text-secondary text-decoration-none filterByKeywordLink">GNNs</a>,
<a href="papers.html?keyword=message passing" target="_blank"
data-tippy-content="See all papers with keyword 'message passing'"
class="text-secondary text-decoration-none filterByKeywordLink">message passing</a>,
<a href="papers.html?keyword=adaptive remeshing" target="_blank"
data-tippy-content="See all papers with keyword 'adaptive remeshing'"
class="text-secondary text-decoration-none filterByKeywordLink">adaptive remeshing</a>,
<a href="papers.html?keyword=encoder-processor-decoder" target="_blank"
data-tippy-content="See all papers with keyword 'encoder-processor-decoder'"
class="text-secondary text-decoration-none filterByKeywordLink">encoder-processor-decoder</a>,
<a href="papers.html?keyword=physical simulation" target="_blank"
data-tippy-content="See all papers with keyword 'physical simulation'"
class="text-secondary text-decoration-none filterByKeywordLink">physical simulation</a>,
<a href="papers.html?keyword=runtime efficiency" target="_blank"
data-tippy-content="See all papers with keyword 'runtime efficiency'"
class="text-secondary text-decoration-none filterByKeywordLink">runtime efficiency</a>,
<a href="papers.html?keyword=neural network simulators" target="_blank"
data-tippy-content="See all papers with keyword 'neural network simulators'"
class="text-secondary text-decoration-none filterByKeywordLink">neural network simulators</a>
</p>
<p class="text-center" style="margin-bottom: 0px;">
<span>Venue: </span>
<a href="papers.html?venue=ICLR" target="_blank" class="text-secondary text-decoration-none">ICLR 2021</a>
</p>
<div class="text-center p-3">
<a class="card-link red-hyper-link" target="_blank" href="https://arxiv.org/pdf/2010.03409.pdf">
Paper
</a>
<a class="card-link red-hyper-link" data-toggle="collapse" role="button" href="#citation">
Citation
</a>
</div>
</div>
<span id="invisible-paper-id" style="display: none;">2</span>
</div>
<div id="citation" class="pp-card m-3 collapse">
<div class="card-body">
<div class="card-text">
<span style="font-size: large; font-weight: bold;">Bibtex:</span>
<span style="white-space: pre-line; position: relative; left: 20px;">
@inproceedings{DBLP:conf/iclr/PfaffFSB21,
author = {Tobias Pfaff and
Meire Fortunato and
Alvaro Sanchez{-}Gonzalez and
Peter W. Battaglia},
title = {Learning Mesh-Based Simulation with Graph Networks},
booktitle = {9th International Conference on Learning Representations, {ICLR} 2021,
Virtual Event, Austria, May 3-7, 2021},
publisher = {OpenReview.net},
year = {2021},
url = {https://openreview.net/forum?id=roNqYL0\_XP},
timestamp = {Wed, 23 Jun 2021 17:36:39 +0200},
biburl = {https://dblp.org/rec/conf/iclr/PfaffFSB21.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
</span>
</div>
<p></p>
</div>
</div>
<div id="details" class="pp-card m-3">
<div class="card-body">
<div class="card-text">
<div id="abstractExample">
<p style="font-weight: bolder; font-size: 25px; text-align: center;">Abstract</p>
Mesh-based simulations are central to modeling complex physical systems in many disciplines across science and engineering. Mesh representations support powerful numerical integration methods and their resolution can be adapted to strike favorable trade-offs between accuracy and efficiency. However, high-dimensional scientific simulations are very expensive to run, and solvers and parameters must often be tuned individually to each system studied. Here we introduce MeshGraphNets, a framework for learning mesh-based simulations using graph neural networks. Our model can be trained to pass messages on a mesh graph and to adapt the mesh discretization during forward simulation. Our results show it can accurately predict the dynamics of a wide range of physical systems, including aerodynamics, structural mechanics, and cloth. The model's adaptivity supports learning resolution-independent dynamics and can scale to more complex state spaces at test time. Our method is also highly efficient, running 1-2 orders of magnitude faster than the simulation on which it is trained. Our approach broadens the range of problems on which neural network simulators can operate and promises to improve the efficiency of complex, scientific modeling tasks.
</div>
</div>
<p></p>
</div>
</div>
<!-- Youtube Video -->
<!-- Project Webpage -->
<!-- Citation Graph -->
<!-- <div class="border-top my-3"></div>
<div class="row p-4">
<div class="col-12" style="text-align: center;">
<span style="font-size: 25px;">Citation Graph</span> <br>
<span>(Double click on nodes to open corresponding papers' pages)</span>
</div>
</div>
<div id="citationGraph" style="width: 100%;
height: 600px;
border: 1px solid lightgray;" ondblclick="openPaperLink()"></div>
<p><span style="font-size: 10px;">*</span> Showing citation graph for papers within our database. Data retrieved from <a href="https://www.semanticscholar.org/search?q=Learning Mesh-Based Simulation with Graph Networks&sort=relevance">Semantic Scholar</a>. For full citation graphs, visit <a href="https://www.connectedpapers.com/search?q=Learning Mesh-Based Simulation with Graph Networks">ConnectedPapers</a>.</p>
<script src="https://cdn.jsdelivr.net/npm/@popperjs/core@2.4.0/dist/umd/popper.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/tippy.js@6/dist/tippy-bundle.umd.min.js"></script>
<script src="static/js/views/paper_detail.js"></script>
<script>
tippy(".filterByAuthorLink");
tippy(".filterByKeywordLink");
const paperID = document.getElementById("invisible-paper-id").innerHTML;
drawCitationGraphAndGenerateCitationCode(paperID);
</script>
-->
</div>
</div>
<!-- body block ends here-->
<!-- Footer -->
<footer class="footer"
style="background-color: #353535; width: 100%; height: 4vh; display: flex; justify-content: center; align-items: center; ">
<!-- no contents needed -->
</footer>
<!-- Code for hash tags -->
<script type="text/javascript">
$(document).ready(function () {
if (window.location.hash !== "") {
$(`a[href="${window.location.hash}"]`).tab("show");
}
$("a[data-toggle='tab']").on("shown.bs.tab", function (e) {
const hash = $(e.target).attr("href");
if (hash.substr(0, 1) === "#") {
const position = $(window).scrollTop();
window.location.replace(`#${hash.substr(1)}`);
$(window).scrollTop(position);
}
});
});
</script>
<!-- footer block ends here -->
</body>
</html>