-
Notifications
You must be signed in to change notification settings - Fork 1
/
xinquan-PINNup.html
699 lines (590 loc) · 26.9 KB
/
xinquan-PINNup.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
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
<style type="text/css">
* { padding: 0; margin: 0;}
.container {
padding: 0px;
margin: 0px;
color: #404040;
font-size: 11pt;
font-family: 'Open Sans', sans-serif;
font-style: normal;
font-weight: 400;
color: #444;
/*background-color: #FFFDFA;*/
/*background: linear-gradient(to right, white, #eee, white);
background: -webkit-linear-gradient( right, white, #eee, white)*/
/*background-image: url("/PaintSwatches_1400x900.jpg");*/
/*font-family: 'textbook', arial, sans-serif;*/
/*margin-left: auto;
margin-right: auto;*/
/*max-width: 2000px;*/
}
body {
background-color: #FFFAFA !important;
background-color: #FAFAFA !important;
}
.section {
padding-top: 20px;
padding-bottom: 20px;
}
.title {
color: #000000;
font-family: 'Roboto Condensed', sans-serif;
padding-top: 10px;
padding-bottom: 25px;
font-size: 26pt;
line-height: 100%;
font-weight: 700;
color: #444;
}
.author {
font-size: 13pt;
}
.section_title {
color: #000000;
font-family: 'Roboto Condensed', sans-serif;
/*padding-top: 10px;*/
/*padding-bottom: 10px;*/
font-size: 20pt;
line-height: 100%;
font-weight: 700;
padding-bottom: 10px;
color: #444;
/*margin-left: auto;
margin-right: auto;
/*line-height: 100%;*/
/*font-weight: 700;*/
}
.shadow {
-webkit-box-shadow: 5px 5px 5px #aaa;
-moz-box-shadow: 5px 5px 5px #aaa;
box-shadow: 5px 5px 5px #aaa;
margin-bottom: 10px;
}
.aff {
margin-bottom: 16px;
/*max-width: 40px;*/
max-height: 30px
}
.img_authors {
position: relative;
max-width: 130px;
max-height: 130px
}
.img_links {
position: relative;
max-width: 70px;
max-height: 70px
}
.text {
position:relative;
line-height:2em;
overflow:hidden;
}
.fadingEffect {
position:absolute;
top:0; bottom:0; right:0;
width:100%;
background:#FFFAFA;
-moz-animation: showHide 5s ease-in alternate infinite; /* Firefox */
-webkit-animation: showHide 5s ease-in alternate infinite; /* Safari and Chrome */
-ms-animation: showHide 5s ease-in alternate infinite; /* IE10 */
-o-animation: showHide 5s ease-in alternate infinite; /* Opera */
animation: showHide 5s ease-in alternate infinite;
}
@-webkit-keyframes showHide { /* Chrome, Safari */
0% {width:100%}
40% {width:0%}
60% {width:0%;}
100% {width:100%;}
}
@-moz-keyframes showHide { /* FF */
0% {width:100%}
40% {width:0%}
60% {width:0%;}
100% {width:100%;}
}
@-ms-keyframes showHide { /* IE10 */
0% {width:100%}
40% {width:0%}
60% {width:0%;}
100% {width:100%;}
}
@-o-keyframes showHide { /* Opera */
0% {width:100%}
40% {width:0%}
60% {width:0%;}
100% {width:100%;}
}
@keyframes showHide {
0% {width:100%}
40% {width:0%}
60% {width:0%;}
100% {width:100%;}
}
#cf {
position:relative;
max-height:301px;
/*width:450px;*/
width:100%;
height:100%;
/*overflow:hidden; */
/*width:inherit;*/
/*max-height: auto;*/
/*margin:0 auto;*/
}
#cf img {
position:absolute ;
max-height:401px;
/*height:100%;*/
/*width:100%;*/
opacity: 0;
left:0;
-webkit-transition: opacity 1s ease-in-out;
-moz-transition: opacity 1s ease-in-out;
-o-transition: opacity 1s ease-in-out;
transition: opacity 1s ease-in-out;
animation-name: cfFadeInOut;
animation-timing-function: ease-in-out;
animation-iteration-count: infinite;
animation-duration: 12s; /* 3images * 4secs */
}
@keyframes cfFadeInOut {
/* fade out */
0% {
opacity:0;
}
5% {
opacity:1;
}
/* fade in */
14.28% {
opacity:1;
}
19.28% {
opacity:0;
}
}
#cf img:nth-of-type(7) {
animation-delay: 24s;
}
#cf img:nth-of-type(6) {
animation-delay: 20s;
}
#cf img:nth-of-type(5) {
animation-delay: 16s;
}
#cf img:nth-of-type(4) {
animation-delay: 12s;
}
#cf img:nth-of-type(3) {
animation-delay: 8s;
}
#cf img:nth-of-type(2) {
animation-delay: 4s;
}
#cf img:nth-of-type(1) {
animation-delay: 0s;
}
.row-centered {
text-align:center;
}
.col-centered {
display:inline-block;
float:none !important;
/* reset the text-align */
text-align:center;
/* inline-block space fix */
margin-right:-4px;
min-width: 170px;
}
html, body {
margin-top: 0px;
margin-bottom: 0px;
margin-left: 10px;
margin-right: 10px;
height: 100%; /* need for iframe height 100% to work */
}
iframe {
box-sizing: border-box; /* make the border size be included in the height */
/*display: block; make them block to fix white space margin */
width: 100%;
}
.hint {
color: #A00;
font-size: 10pt;
margin-bottom: 10px;
}
.img-process {
/*width: 100%;*/
margin: 5px;
max-height: 200px;
-webkit-transition: opacity 0.2s ease-in-out;
-moz-transition: opacity 0.2s ease-in-out;
-ms-transition: opacity 0.2s ease-in-out;
-o-transition: opacity 0.2s ease-in-out;
transition: opacity 0.2s ease-in-out;
opacity: 1;
}
.img-process:hover{
cursor: pointer;
opacity: 0.6;
}
.centeredd {
position: fixed;
top: 50%;
left: 50%;
transform: translate(-50%, -50%);
-webkit-transform: translate(-50%, -50%);
-moz-transform: translate(-50%, -50%);
-o-transform: translate(-50%, -50%);
-ms-transform: translate(-50%, -50%);
}
</style>
<!doctype html>
<html lang="en">
<head>
<!-- <meta name="viewport" content="width=device-width, initial-scale=1.0"> -->
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<title>SWAG Publication</title>
<meta property="og:image" content="https://swagroup.kaust.edu.sa/images/default-source/homepageimages/swag---new-logo.jpg?sfvrsn=e3c9e798_2&MaxWidth=250&MaxHeight=107&ScaleUp=false&Quality=High&Method=ResizeFitToAreaArguments&Signature=71FD4BD748A81E57E67CB9A68CC991B285D2C9ED"/>
<meta property="og:title" content="Seismic Wave Analysis - Publication item" />
<!-- Latest compiled and minified CSS -->
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/css/bootstrap.min.css" integrity="sha384-BVYiiSIFeK1dGmJRAkycuHAHRg32OmUcww7on3RYdg4Va+PmSTsz/K68vbdEjh4u" crossorigin="anonymous">
<!-- Optional theme -->
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/css/bootstrap-theme.min.css" integrity="sha384-rHyoN1iRsVXV4nD0JutlnGaslCJuC7uwjduW9SVrLvRYooPp2bWYgmgJQIXwl/Sp" crossorigin="anonymous"><!-- <link rel="stylesheet" href="../css/bootstrap-theme.min.css"> -->
<!-- Google fonts -->
<!-- <link href="../css/google-fonts.css" rel="stylesheet" type="text/css"> -->
<link href="https://fonts.googleapis.com/css?family=Roboto+Condensed:400,700|Open+Sans:300italic,400italic,600italic,400,700,300,600" rel="stylesheet" type="text/css">
<!-- Photoswipe -->
<link href="https://cdnjs.cloudflare.com/ajax/libs/photoswipe/4.1.2/photoswipe.css" rel="stylesheet">
<link href="https://cdnjs.cloudflare.com/ajax/libs/photoswipe/4.1.2/default-skin/default-skin.css" rel="stylesheet">
<script type="text/x-mathjax-config">
MathJax.Hub.Config({
tex2jax: {inlineMath: [['$','$'], ['\\(','\\)']]},
TeX: { equationNumbers: { autoNumber: "AMS" } },
});
</script>
<script type="text/javascript" async
src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.2/MathJax.js?config=TeX-MML-AM_CHTML">
</script>
<script>
function resizeIframe(obj) {
obj.style.height = 0;
obj.style.height = (obj.contentWindow.document.body.scrollHeight + 5) + 'px';
}
</script>
<script language="JavaScript">
function resize_iframe()
{
obj = document.getElementById("myiframe")
obj.style.height = 0;
obj.style.height = (obj.contentWindow.document.body.scrollHeight) + 'px';
document.getElementById("cf").style.height = document.getElementById("lastimage").height + 8;
}
window.onresize=resize_iframe;
</script>
</head>
<body>
<script src="zepto.min.js"></script>
<script>
window.addEventListener('message', function(event) {
if(height = event.data['height']) {
$('iframe').css('height', height + 'px')
}
});
// var resizeEvent = new Event('resize');
// window.dispatchEvent(resizeEvent);
</script>
<!-- ====================================================== -->
<!-- ===================== Header ========================= -->
<!-- ====================================================== -->
<center>
<div class="row section" style="margin-bottom: -24px">
<div class="col-sm-5" style="margin-left: 1px">
<img src="assets/images/KAUST_logo.png" class="img_links img-responsive">
</div>
<div class="col-sm-5" style="margin-right: 1px">
<img src="assets/images/SWAG_logo.png" class="img_links img-responsive">
</div>
</div>
</center>
<hr>
<div class="container">
<!-- ====================================================== -->
<!-- =========== To ADD your Research ===================== -->
<!-- ========== START EDITING FROM HERE =================== -->
<!-- ====================================================== -->
<!-- ====================================================== -->
<!-- ===================== TITLE ========================== -->
<!-- ====================================================== -->
<center>
<div class="title">
PINNup: Robust neural network wavefield solutions using
frequency upscaling and neuron splitting
</div>
</center>
<center>
<div class="row row-eq-height" style="margin-bottom: -10px">
<div class="col-sm-12 col-sm-offset-2">
<div class="col-sm-4" style="margin-bottom: 10px">
<a href="https://swagroup.kaust.edu.sa/people/detail/xinquan-huang">
<img src="assets/authors/XinquanHuang.jpg" class="img_authors img-circle">
<div class="author"><span style="font-weight: 600">Xinquan Huang</span><br>KAUST</div>
</a>
</div>
<div class="col-sm-4" style="margin-bottom: 10px">
<a href="https://swagroup.kaust.edu.sa/people/detail/tariq-alkhalifah">
<img src="assets/authors/TariqAlkhalifah.png" class="shadow img_authors img-circle img-responsive">
<div class="author"><span style="font-weight: 600">Tariq Alkhalifah</span><br>KAUST</div>
</a>
</div>
</div>
</div>
</center>
<!-- ====================================================== -->
<!-- ===================== LINKS ========================== -->
<!-- ====================================================== -->
<center>
<div class="row section" style="margin-bottom: -12px">
<div class="col-sm-10 col-sm-offset-1">
<!-- <div class="col-sm-12 center"> -->
<div class="col-sm-3" style="margin-bottom: 12px">
<!-- <div class="col-sm-12"> -->
<a href="https://repository.kaust.edu.sa/handle/10754/672118">
<img src="assets/images/pdf.svg" class="img_links img-responsive">
<div class="author">Paper</div>
</a>
<!-- </div> -->
</div>
<div class="col-sm-3" style="margin-bottom: 12px">
<!-- <div class="col-sm-12"> -->
<a href="">
<img src="assets/images/pdf.svg" class="img_links img-responsive" >
<div class="author">Supplementary Material</div>
</a>
<!-- </div> -->
</div>
<div class="col-sm-3" style="margin-bottom: 12px">
<!-- <div class="col-sm-12"> -->
<a href="https://doi.org/10.1029%2F2021jb023703">
<img src="assets/images/pdf.svg" class="img_links img-responsive" >
<div class="author">Paper (journal)</div>
</a>
<!-- </div> -->
</div>
<div class="col-sm-3" style="margin-bottom: 12px">
<!-- <div class="col-sm-12"> -->
<a href="https://github.com/swag-kaust/PINNup" >
<img src="assets/images/github.svg" class="img_links img-responsive">
<div class="author">Code</div>
</a>
<!-- </div> -->
</div>
<!-- </div> -->
</div>
</div>
</center>
<!-- ====================================================== -->
<!-- ===================== TEASER ============== -->
<!-- ====================================================== -->
<br>
<!-- <center> -->
<div class="row row-eq-height" style="margin-bottom: -10px">
<div class="col-sm-12 col-sm-offset-0">
<div class="col-sm-6" style="margin-bottom: 10px">
<br>
<div class="section_title">Highlights</div>
<ol style="list-style-type: circle">
<li>We propose a novel physics-informed neural network framework using frequency upscaling and neuron splitting, yielding efficient, highly accurate, high-frequency wavefield solutions. </li>
<li>We propose an empirical formula that relates neuron splitting to frequency upscaling that allows for high accuracy and fast convergence.</li>
<li>It offers a compressed representation of wavefield solution, allowing for easy storage, and instant access of wavefield on dynamic grids.</li>
</ol>
</div>
<div class="col-sm-6" style="margin-bottom: 10px">
<div id="cf">
<center>
<img src="assets/xinquan-PINNup/8hz.png" class="img-responsive" id="firstimage"/>
<img src="assets/xinquan-PINNup/16hz.png" class="img-responsive"/>
<img src="assets/xinquan-PINNup/32hz.png" class="img-responsive" id="lastimage"/>
</center>
</div>
</div>
</div>
</div>
<!-- </center> -->
<!-- <center>
<div class="row section center row-eq-height">
<div class="col center col-sm-8 col-sm-offset-2">
<div id="cf">
<center>
<img src="assets/birnie-self-supervised-networks-for-random-noise-suppression/Teaser.jpg" class="img-responsive" id="firstimage"/>
<img src="assets/birnie-self-supervised-networks-for-random-noise-suppression/Teaser.jpg" class="img-responsive"/>
<img src="assets/birnie-self-supervised-networks-for-random-noise-suppression/Teaser.jpg" class="img-responsive"/>
<img src="assets/birnie-self-supervised-networks-for-random-noise-suppression/Teaser.jpg" class="img-responsive" id="lastimage"/>
</center>
</div>
<span style="font-size:14px"><i></i>
</span>
</div>
</div>
</center> -->
<hr>
<!-- ====================================================== -->
<!-- ===================== Abstract ======================= -->
<!-- ====================================================== -->
<div class="row section">
<div class="section_title">Abstract</div>
<p>
Seismic wave-equation based methods, for example, full waveform inversion, are currently used to illuminate the interior of Earth. Solving for the frequency-domain scattered wavefield via physics-informed neural network (PINN) has great potential in increasing the flexibility and reducing the computational cost of seismic modeling and inversion. However, when dealing with high-frequency wavefields using PINN, its accuracy and training cost limit its application. Thus, we propose a novel implementation of PINN using frequency upscaling and neuron splitting, which allows the neural network model to grow in size as we increase the frequency while leveraging the information from the pre-trained model for lower-frequency wavefields, resulting in fast convergence to highly accurate wavefield solutions. Numerical results show that, compared to the commonly used PINN with random initialization, the proposed PINN exhibits notable superiority in terms of convergence and accuracy and can achieve neuron based high-frequency wavefield solutions with a shallow model.
</p>
</div>
<!-- ====================================================== -->
<!-- ================== Theory ============================ -->
<!-- ====================================================== -->
<div class="row section">
<div class="section_title">Theory</div>
Similar to FWI, our NN is first trained for low-frequency wavefields, and gradually optimized for high-frequency wavefields using the lower frequency NN parameters to initialize the model. The information contained in the NN for low-frequency wavefields is beneficial for higher-frequency wavefield training, leading to faster convergence and better prediction than training from scratch. From the prospective of deep learning, our source domain (low-frequency wavefield) and target domain (a higher-frequency wavefield) are inherently related through the kinematic properties for a given velocity model and source location. However, higher frequency wavefields are dynamically more complex, and thus, will require larger neural network models that can represent the complex features. The workflow of the proposed method is illustrated in Figure 1. After we train the neural network to predict low-frequency wavefields using a small network, we increase the size of the network through neural splitting and then use it to learn higher frequency representation. Thus, we can benefit from the NN training experience at low frequency to help us converge faster at high frequencies.
<p>
<center>
<p>
<img src="assets/xinquan-PINNup/figure1.png" width="800" class="img-responsive">
<!-- <br /> <em>Fig. 1</em> Schematic illustration of the Noise2Void denoising procedure. -->
</p>
</center>
<br>
<p>
As mentioned before, we want to leverage the information from trained models, while increasing the convergence speed of larger NNs needed for higher frequency representations. The concept of neuron splitting allows us to increase the network size without effecting its output. The splitting process of the neurons in a hidden layer involves duplicating the weights coming into the neuron to all of its off springs, while dividing the weights connecting the neuron with the next layer by the number of offsprings. In the case of splitting all the neurons, the weights and biases are given by the following formulas:
\begin{equation}
\label{equ:inputl}
\mathbf{W}^{(1)}_{split} = \begin{bmatrix} \mathbf{W}^{(1)} \dots \mathbf{W}^{(1)} \end{bmatrix}^T, \mathbf{b}^{(1)}_{split} = \begin{bmatrix} \mathbf{b}^{(1)}\dots\mathbf{b}^{(1)}\end{bmatrix}^T,
\end{equation}
\begin{equation}
\label{equ:hidl}
\mathbf{W}^{(\ell)}_{split} = \frac{1}{n}\begin{bmatrix} \mathbf{W}^{(\ell)} \dots \mathbf{W}^{(\ell)} \\
\vdots \\
\mathbf{W}^{(\ell)} \dots \mathbf{W}^{(\ell)}
\end{bmatrix}, \mathbf{b}^{(\ell)}_{split} = \begin{bmatrix} \mathbf{b}^{(\ell)}\dots\mathbf{b}^{(\ell)}\end{bmatrix}^T,
\end{equation}
\begin{equation}
\label{equ:outl}
\mathbf{W}^{(L)}_{split} = \frac{1}{n}\begin{bmatrix} \mathbf{W}^{(L)} \dots \mathbf{W}^{(L)} \end{bmatrix}, \mathbf{b}^{(L)}_{split} = \mathbf{b}^{(L)}.
\end{equation}
where the size of the vector and the number of columns in the matrix are equal to $n$.
<!-- ====================================================== -->
<!-- ================== Results =========================== -->
<!-- ====================================================== -->
<div class="row section">
<div class="section_title">Results</div>
<p>
Even though the representations of wavefield via NN at every frequency may include some errors, those errors, as we saw, are generally small, and we managed to capture the most important features of the wavefield. To further highlight the importance of our approach for PINN, we refer the reader to Sitzmann et al. (2020) in which they use a sine activation function to predict a 3.2 Hz wavefield corresponding to the Helmholtz equation, and to do so they needed a network of 5 layers with 256 neurons per layer, and that prediction was for a single source, not a Green's function.
</p>
<p>
From the experiments, we observe that with our proposed method, we can easily predict highly accurate multi-frequency wavefield solutions at any location in the domain of interest corresponding to any source location on the surface (because the input of our network includes the coordinates of source on the surface and arbitrary space coordinates). In other words, no interpolation is needed. Such a continuous wavefield representation is stored in a much more compressed form even for the 32 Hz representation.
</p>
<p>
Experiments on the Overthrust model demonstrate that the proposed method outperforms the vanilla PINN in accuracy and efficiency.
</p>
<p>
<center>
<img src="assets/xinquan-PINNup/layer.001.png" class="img-responsive">
<img src="assets/xinquan-PINNup/ove.001.png" class="img-responsive">
</div>
<!--<center>
<p>
<img src="assets/birnie-self-supervised-networks-for-random-noise-suppression/Results.png" class="img-responsive">-->
<!-- <br /> <em>Fig. 1</em> Schematic illustration of the Noise2Void denoising procedure. -->
<!--</p>
</center>-->
<hr>
<!-- ====================================================== -->
<!-- ================== References ======================== -->
<!-- ====================================================== -->
<div class="row section">
<div class="section_title">Citation</div>
<p>
If you found the paper useful, please cite it via:<br />
</p>
<div class="highlight wb">
<pre class="chroma">
<code>
Huang, X., & Alkhalifah, T. (2022). PINNup: Robust neural network wavefield solutions using frequency upscaling and neuron splitting. Journal of Geophysical Research: Solid Earth, 127, e2021JB023703. https://doi.org/10.1029/2021JB023703
</code>
</pre>
</div>
<div class="highlight">
<pre class="chroma">
<code>
@article{Huang2022PINNup,
doi = {10.1029/2021jb023703},
url = {https://doi.org/10.1029%2F2021jb023703},
year = 2022,
month = {jun},
volume={127},
issue={6},
publisher = {American Geophysical Union ({AGU})},
author = {Xinquan Huang and Tariq Alkhalifah},
title = {{PINNup}: Robust neural network wavefield solutions using frequency upscaling and neuron splitting},
journal = {Journal of Geophysical Research: Solid Earth} }
</code>
</pre>
</div>
<p>
<!-- See <a href="">paper</a> and <a href="">supplementary material</a> for details. -->
</p>
</div>
</div>
<!-- ====================================================== -->
<!-- =========== To ADD your Research ===================== -->
<!-- ============== STOP EDITING HERE ===================== -->
<!-- ====================================================== -->
<!-- ====================================================== -->
<!-- ===================== Footer ========================= -->
<!-- ====================================================== -->
<center>
<div class="container">
</div><div class="panel-footer">
<p class="footer-company-name">All Rights Reserved. © 2022
<a href="https://swagroup.kaust.edu.sa/">Seismic Wave Analysis Group</a>
</p>
</div>
</div><!-- end container -->
</center>
<!-- Optional JavaScript -->
<!-- jQuery first, then Popper.js, then Bootstrap JS -->
<script src="https://code.jquery.com/jquery-3.2.1.slim.min.js" integrity="sha384-KJ3o2DKtIkvYIK3UENzmM7KCkRr/rE9/Qpg6aAZGJwFDMVNA/GpGFF93hXpG5KkN" crossorigin="anonymous"></script>
<!-- <script src="https://cdnjs.cloudflare.com/ajax/libs/popper.js/1.12.3/umd/popper.min.js" integrity="sha384-vFJXuSJphROIrBnz7yo7oB41mKfc8JzQZiCq4NCceLEaO4IHwicKwpJf9c9IpFgh" crossorigin="anonymous"></script> -->
<!-- <script src="https://maxcdn.bootstrapcdn.com/bootstrap/4.0.0-beta.2/js/bootstrap.min.js" integrity="sha384-alpBpkh1PFOepccYVYDB4do5UnbKysX5WZXm3XxPqe5iKTfUKjNkCk9SaVuEZflJ" crossorigin="anonymous"></script> -->
<script src="https://cdn.jsdelivr.net/npm/@popperjs/core@2.10.2/dist/umd/popper.min.js" integrity="sha384-7+zCNj/IqJ95wo16oMtfsKbZ9ccEh31eOz1HGyDuCQ6wgnyJNSYdrPa03rtR1zdB" crossorigin="anonymous"></script>
<script src="https://cdn.jsdelivr.net/npm/bootstrap@5.1.3/dist/js/bootstrap.min.js" integrity="sha384-QJHtvGhmr9XOIpI6YVutG+2QOK9T+ZnN4kzFN1RtK3zEFEIsxhlmWl5/YESvpZ13" crossorigin="anonymous"></script>
<script type="text/javascript">
function change_height_cf(el) {
if (el.height > 0) {
document.getElementById("cf").style.height = el.height + 8;
};
};
$(document).ready(function(){
$('#firstimage').on('load', function(){
console.log(2);
change_height_cf(document.getElementById("firstimage"));
});
if ($('#firstimage')[0].complete) {
console.log(3);
change_height_cf(document.getElementById("firstimage"));
};
$('#lastimage').on('load', function(){
console.log(2);
change_height_cf(document.getElementById("lastimage"));
});
if ($('#lastimage')[0].complete) {
console.log(3);
change_height_cf(document.getElementById("lastimage"));
};
});
</script>
<!-- photoswipe -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/photoswipe/4.1.2/photoswipe.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/photoswipe/4.1.2/photoswipe-ui-default.min.js"></script>
<script src="img-gallery.js"></script>
</body>
</html>