-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
1236 lines (864 loc) · 41.1 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
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
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="zh-CN">
<head>
<meta charset="utf-8" />
<meta name="keywords" content="课程, code, 生活, 分享" />
<meta
name="viewport"
content="width=device-width, initial-scale=1, maximum-scale=1"
/>
<title> Hy's Treasure house</title>
<meta name="generator" content="hexo-theme-ayer">
<link rel="shortcut icon" href="/favicon.ico" />
<link rel="stylesheet" href="/dist/main.css">
<link rel="stylesheet" href="/css/fonts/remixicon.css">
<link rel="stylesheet" href="/css/custom.css">
<script src="https://cdn.staticfile.net/pace/1.2.4/pace.min.js"></script>
<link
rel="stylesheet"
href="https://cdn.jsdelivr.net/npm/@sweetalert2/theme-bulma@5.0.1/bulma.min.css"
/>
<script src="https://cdn.jsdelivr.net/npm/sweetalert2@11.0.19/dist/sweetalert2.min.js"></script>
<!-- mermaid -->
<style>
.swal2-styled.swal2-confirm {
font-size: 1.6rem;
}
</style>
<link rel="stylesheet" href="\assets\css\APlayer.min.css" class="aplayer-style-marker">
<script src="\assets\js\APlayer.min.js" class="aplayer-script-marker"></script>
</head>
</html>
</html>
<!--浏览器搞笑标题-->
<script type="text/javascript" src="\js\FunnyTitle.js"></script>
<body>
<div id="app">
<canvas width="1777" height="841"
style="position: fixed; left: 0px; top: 0px; z-index: 99999; pointer-events: none;"></canvas>
<main class="content on">
<section class="cover">
<a class="forkMe" href="https://github.com/hydrogenhy"
target="_blank"><img width="149" height="149" src="/images/forkme.png"
class="attachment-full size-full" alt="Fork me on GitHub" data-recalc-dims="1"></a>
<div class="cover-frame">
<div class="bg-box">
<img src="/images/cover2.jpg" alt="image frame" />
</div>
<div class="cover-inner text-center text-white">
<h1><a href="/">Hy's Treasure house</a></h1>
<div id="subtitle-box">
<span id="subtitle"></span>
</div>
<div>
</div>
</div>
</div>
<div class="cover-learn-more">
<a href="javascript:void(0)" class="anchor"><i class="ri-arrow-down-line"></i></a>
</div>
</section>
<script src="https://cdn.staticfile.net/typed.js/2.0.12/typed.min.js"></script>
<!-- Subtitle -->
<script>
try {
var typed = new Typed("#subtitle", {
strings: ['竹杖芒鞋轻胜马，谁怕？一蓑烟雨任平生。', '', ''],
startDelay: 0,
typeSpeed: 200,
loop: true,
backSpeed: 100,
showCursor: true
});
} catch (err) {
console.log(err)
}
</script>
<div id="main">
<section class="outer">
<ul class="ads">
<li>
<a href="">
<img src="" width="0" alt="">
</a>
</li>
<li>
<a href="">
<img src="" width="0" alt="">
</a>
</li>
</ul>
<div class="notice" style="margin-top:50px">
<i class="ri-heart-fill"></i>
<div class="notice-content" id="broad"></div>
</div>
<script type="text/javascript">
fetch('https://v1.hitokoto.cn')
.then(response => response.json())
.then(data => {
document.getElementById("broad").innerHTML = data.hitokoto;
})
.catch(console.error)
</script>
<style>
.notice {
padding: 20px;
border: 1px dashed #e6e6e6;
color: #969696;
position: relative;
display: inline-block;
width: 100%;
background: #fbfbfb50;
border-radius: 10px;
}
.notice i {
float: left;
color: #999;
font-size: 16px;
padding-right: 10px;
vertical-align: middle;
margin-top: -2px;
}
.notice-content {
display: initial;
vertical-align: middle;
}
</style>
<article class="articles">
<article
id="post-Hi-很高兴见到你"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2023/07/17/Hi-%E5%BE%88%E9%AB%98%E5%85%B4%E8%A7%81%E5%88%B0%E4%BD%A0/"
>Hi, 很高兴见到你
<i class="article-topping">置顶</i>
</a>
</h2>
</header>
<div class="article-meta">
<a href="/2023/07/17/Hi-%E5%BE%88%E9%AB%98%E5%85%B4%E8%A7%81%E5%88%B0%E4%BD%A0/" class="article-date">
<time datetime="2023-07-17T07:37:03.000Z" itemprop="datePublished">2023-07-17</time>
</a>
</div>
<div class="article-entry" itemprop="articleBody">
<div id="aplayer-tVoIbLMN" class="aplayer aplayer-tag-marker" style="margin-bottom: 20px;">
<pre class="aplayer-lrc-content"></pre>
</div>
<script>
var ap = new APlayer({
element: document.getElementById("aplayer-tVoIbLMN"),
narrow: false,
autoplay: false,
showlrc: false,
music: {
title: "日落大道",
author: "梁博",
url: "/music/日落大道.mp3",
pic: "/music/日落大道.jpg",
lrc: ""
}
});
window.aplayers || (window.aplayers = []);
window.aplayers.push(ap);
</script>
<center><strong><font size="5">欢迎来到我的博客！</font></strong></center>
<p>  在这里我将分享我的<strong>学习心得</strong>、<strong>some projects</strong>、<strong>吉他谱</strong>和<strong>一些个人想法</strong>等等。希望能够对你有所帮助。</p>
<p>  创建博客博客的初衷，一是希望有个<strong>个人空间</strong>来存储我的经历，博客就是很好的选择；二是我希望我的心得能够让更多人看见，<strong>帮助一些人</strong>少走一些弯路——在学习过程中我被各路“开源”大佬的文章启发，因此我也想将我的思考“开源”出去。</p>
<p>  整个博客使用<code>hexo</code>搭建，主题是<code>ayer</code>。这两个链接我放在了“<strong>友链</strong>”中供大家参考。此外也运用了如播放器、搜索等功能，大家可自行去探索<code>hexo</code>的诸多功能。</p>
<p> <strong> 那么，enjoy！！！; -)</strong></p>
<p><del>所以你是被谁骗到这里的？</del></p>
<p><strong>既然来了，Why not leave a comment ? （要先点进此博客）</strong> </p>
<p align="right"><strong>2023.07.17</strong></p>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/About/" rel="tag">About</a></li></ul>
</footer>
</div>
</article>
<article
id="post-SIB"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/09/12/SIB/"
>《Recognizing Predictive Substructures with Subgraph Information Bottleneck》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/09/12/SIB/" class="article-date">
<time datetime="2024-09-12T11:27:33.000Z" itemprop="datePublished">2024-09-12</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://ieeexplore.ieee.org/document/9537601">click here</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>本文是TPAMI 2024中的一篇工作。</p>
<p>这个论文其实21年就已经发在网上了，24年这个论文和21年的变化基本不大，目测只加了一个图。</p>
<p>这篇论文目的不是靠信息瓶颈找到一个潜在表示，而是抽象出一个子图结构。
<a class="article-more-link" href="/2024/09/12/SIB/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-over-squashing and bottlenecks via curvature"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/18/over-squashing%20and%20bottlenecks%20via%20curvature/"
>《Understanding over-squashing and bottlenecks on graphs via curvature》论文浅析</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/18/over-squashing%20and%20bottlenecks%20via%20curvature/" class="article-date">
<time datetime="2024-08-18T06:32:33.000Z" itemprop="datePublished">2024-08-18</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://openreview.net/pdf?id=7UmjRGzp-A">click here</a></p>
<h3 id="论文简述"><a href="#论文简述" class="headerlink" title="论文简述"></a>论文简述</h3><p>本文是ICLR 2022中的一篇高分工作（openreview：10/8/8/8）。</p>
<p>本文从理论上提出了一种Ricci曲率计算方法，并根据对其的分析指出重要结论：<strong>负曲率是导致瓶颈的原因</strong>。而后将其与Cheeger constant联系，说明了他们一定程度上是比较相似的，后期也使用Cheeger constant来说明随机游走重建方法的劣势。
<a class="article-more-link" href="/2024/08/18/over-squashing%20and%20bottlenecks%20via%20curvature/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/Over-Squashing/" rel="tag">Over-Squashing</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-Residual Entropy"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/18/Residual%20Entropy/"
>《Residual Entropy-based Graph Generative Algorithms》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/18/Residual%20Entropy/" class="article-date">
<time datetime="2024-08-18T03:52:33.000Z" itemprop="datePublished">2024-08-18</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://dl.acm.org/doi/10.5555/3535850.3535942">click here</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>本文是AAMAS 2022中的一篇工作。</p>
<p>本篇文章的核心工作为：<strong>设计了一种通过原图构造防御图的算法（非神经网络）</strong>。其意义在于通过对原始graph进行加边处理使得其结构信息更加显著，从而帮助模型更好的学习到structure信息。</p>
<p>文章通过1-dimensional structure entropy和2-dimensional structure entropy定义出残差熵（Residual Entropy），这个熵反映了图结构所携带的信息——本文希望最大化这个量。</p>
<p>文章通过残差熵的定义推导出使得残差熵增大的条件，进而根据这个条件对图进行加边处理，旨在提高图的残差熵，从而得到<strong>防御图</strong>。
<a class="article-more-link" href="/2024/08/18/Residual%20Entropy/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-desktop"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/16/desktop/"
>一亩三分地——宿舍桌面v2.0版</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/16/desktop/" class="article-date">
<time datetime="2024-08-16T15:24:20.000Z" itemprop="datePublished">2024-08-16</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%9D%82/">杂</a> / <a class="article-category-link" href="/categories/%E6%9D%82/%E7%94%9F%E6%B4%BB%E5%88%86%E4%BA%AB/">生活分享</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>hi，陆陆续续，从上次装修宿舍以来也已经过了一年了吧。</p>
<p>这期间又经历了很多事，大学从 50% 到 75%，有多少难熬的夜晚，有多少内心的悸动，但都经历过了。岁月留痕，一件物品，一段故事。
<a class="article-more-link" href="/2024/08/16/desktop/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E6%9D%82/" rel="tag">杂</a></li></ul>
</footer>
</div>
</article>
<article
id="post-Edge Entropy"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/11/Edge%20Entropy/"
>《Edge Entropy as an Indicator of the Effectiveness of GNNs over CNNs for Node Classification》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/11/Edge%20Entropy/" class="article-date">
<time datetime="2024-08-11T05:54:20.000Z" itemprop="datePublished">2024-08-11</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://ieeexplore.ieee.org/document/9443451">click here</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>这是CMU在2021 Asilomar Conference上的一篇工作。</p>
<p>GNN与CNN的区别在于，GNN引入了图的结构信息。从实验结果来看，有效的图结构训练结果要远好于简单的单位阵或随机矩阵当做图结构的结果：
<a class="article-more-link" href="/2024/08/11/Edge%20Entropy/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-Delaunay Graph"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/09/Delaunay%20Graph/"
>《Delaunay Graph：Addressing Over-Squashing and Over-Smoothing Using Delaunay Triangulation》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/09/Delaunay%20Graph/" class="article-date">
<time datetime="2024-08-09T08:16:20.000Z" itemprop="datePublished">2024-08-09</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://openreview.net/pdf?id=uyhjKoaIQa">click here</a></p>
<p>code：<a target="_blank" rel="noopener" href="https://github.com/Hugo-Attali/Delaunay-Rewiring/blob/main">Delaunay Graph(github)</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>本文是来自2024ICML的论文。</p>
<p>图神经网络在计算节点的特征时采用了聚合的思路，但随着gnn层数的增长，会面对如下两个问题：</p>
<ul>
<li><p><strong>Over-smoothing</strong>：如果GNN层很深的话，随着聚合节点的不断扩散，任意两个节点所共享的邻居就会非常多（感受野重叠），导致这两个节点的嵌入非常相似。
<a class="article-more-link" href="/2024/08/09/Delaunay%20Graph/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/Over-Smoothing/" rel="tag">Over-Smoothing</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/Over-Squashing/" rel="tag">Over-Squashing</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-rethinking CEL"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/02/rethinking%20CEL/"
>《Rethinking Independent Cross-Entropy Loss For Graph-Structured Data》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/02/rethinking%20CEL/" class="article-date">
<time datetime="2024-08-02T08:15:20.000Z" itemprop="datePublished">2024-08-02</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://proceedings.mlr.press/v235/zhang24bt.html">click here</a></p>
<p>code：<a target="_blank" rel="noopener" href="https://github.com/MR9812/Joint-Cluster-Supervised-Learning">Joint-Cluster(github)</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>本文是来自2024ICML的论文。</p>
<p> 对于节点的表示与预测，会存在如下问题：</p>
<ul>
<li>由独立条件分布 $P(y_i|z_i)$ 和 $CE(y_i, P(y_i|z_i))$ 所导致的：<ul>
<li>过拟合</li>
<li>容易受到adversarial attacks的影响
<a class="article-more-link" href="/2024/08/02/rethinking%20CEL/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-GeMax"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/08/01/GeMax/"
>《Learning Graph Representation via Graph Entropy Maximization》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/08/01/GeMax/" class="article-date">
<time datetime="2024-08-01T01:30:20.000Z" itemprop="datePublished">2024-08-01</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://proceedings.mlr.press/v235/sun24i.html">click here</a></p>
<p>code：<a target="_blank" rel="noopener" href="https://github.com/MathAdventurer/GeMax">GeMax(github)</a></p>
<h3 id="论文任务与个人想法"><a href="#论文任务与个人想法" class="headerlink" title="论文任务与个人想法"></a>论文任务与个人想法</h3><p>本工作来自于2024ICML。</p>
<p>文章的任务是图表示学习，其不仅对每一张图上的<strong>每个节点得到编码</strong>，更在图整体上得到关于<strong>图的编码</strong>。本文提出了一种利<strong>用正交表示进行图熵最大化</strong>的近似方法。</p>
<p>整篇文章读下来，有让我眼前一亮的感觉，是让我觉得蛮有意思的一篇文章。</p>
<p>文章运用了其他文献未运用的，基于<strong>独立集</strong>表示的熵的表示方法。
<a class="article-more-link" href="/2024/08/01/GeMax/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
<article
id="post-DGIB"
class="article article-type-post"
itemscope
itemprop="blogPost"
data-scroll-reveal
>
<div class="article-inner">
<header class="article-header">
<h2 itemprop="name">
<a class="article-title" href="/2024/07/26/DGIB/"
>《Dynamic Graph Information Bottleneck》论文解读</a>
</h2>
</header>
<div class="article-meta">
<a href="/2024/07/26/DGIB/" class="article-date">
<time datetime="2024-07-26T12:03:13.000Z" itemprop="datePublished">2024-07-26</time>
</a>
<div class="article-category">
<a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/">深度学习</a> / <a class="article-category-link" href="/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/">图神经网络</a>
</div>
</div>
<div class="article-entry" itemprop="articleBody">
<p>论文传送门：<a target="_blank" rel="noopener" href="https://arxiv.org/abs/2402.06716">click here</a></p>
<p>Code：<a target="_blank" rel="noopener" href="https://github.com/RingBDStack/DGIB/tree/main">DGIB(github.com)</a></p>
<h3 id="论文动机"><a href="#论文动机" class="headerlink" title="论文动机"></a>论文动机</h3><p>本篇论文是2024WWW上的论文。</p>
<p>这篇论文将<strong>信息瓶颈理论</strong>创新性地应用于动态图神经网络中，遇上一篇论文类似，文章核心围绕着信息瓶颈理论的——minimal sufficient 特点展开。但动态图具有时间特征，因此在时间特征上的编码一致性是重要的，针对此问题，本文将minimal - sufficient进行扩展，<strong>得到Minimal-Sufficient-Consensual (MSC)原则</strong>。</p>
<a class="article-more-link" href="/2024/07/26/DGIB/"
>阅读更多...</a
>
<!-- reward -->
</div>
<!-- copyright -->
<footer class="article-footer">
<ul class="article-tag-list" itemprop="keywords"><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E4%BF%A1%E6%81%AF%E8%AE%BA/" rel="tag">信息论</a></li><li class="article-tag-list-item"><a class="article-tag-list-link" href="/tags/%E5%9B%BE%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/" rel="tag">图神经网络</a></li></ul>
</footer>
</div>
</article>
</article>
<nav class="page-nav">
<span class="page-number current">1</span><a class="page-number" href="/page/2/">2</a><a class="page-number" href="/page/3/">3</a><a class="page-number" href="/page/4/">4</a><a class="extend next" rel="next" href="/page/2/">下一页</a>
</nav>
</section>
</div>
<footer class="footer">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Runtime Display</title>
<style>
/* 将运行时间的容器居中并加粗 */
#runtime_span {
display: block;
text-align: center; /* 居中 */
font-weight: bold; /* 加粗 */
font-size: 18px; /* 调整字体大小 */
}
</style>
</head>
<span id="runtime_span"></span>
<script type="text/javascript">
function show_runtime() {
window.setTimeout("show_runtime()", 1000);
X = new Date("2023/07/16 00:22:00"); // 网站开始运行的日期和时间
Y = new Date(); // 当前日期和时间
T = (Y.getTime() - X.getTime()); // 网站运行的总毫秒数
M = 24 * 60 * 60 * 1000; // 一天的毫秒数
a = T / M; // 总天数
A = Math.floor(a); // 总天数的整数部分
b = (a - A) * 24; // 总小时数
B = Math.floor(b); // 总小时数的整数部分
c = (b - B) * 60; // 总分钟数
C = Math.floor((b - B) * 60); // 总分钟数的整数部分
D = Math.floor((c - C) * 60); // 总秒数
runtime_span.innerHTML = "⏱️勉强运行 <span class='runtime'>" + A + " 天 " + B + " 小时 " + C + " 分 " + D + " 秒</span>";
}
show_runtime();
</script>
<div class="outer">
<ul>
<li>
Copyrights ©
2023-2024
<i class="ri-heart-fill heart_icon"></i> Lucas
</li>
</ul>
<ul>
<li>
</li>
</ul>
<ul>
<li>
<span>
<span><i class="ri-user-3-fill"></i>访问人数:<span id="busuanzi_value_site_uv"></span></span>
<span class="division">|</span>
<span><i class="ri-eye-fill"></i>浏览次数:<span id="busuanzi_value_page_pv"></span></span>
</span>
</li>
</ul>
<ul>
</ul>
<ul>
</ul>
<ul>
<li>
<!-- cnzz统计 -->
</li>
</ul>
</div>
</footer>
</main>
<div class="float_btns">
<div class="totop" id="totop">
<i class="ri-arrow-up-line"></i>
</div>
<div class="todark" id="todark">
<i class="ri-moon-line"></i>
</div>
</div>
<aside class="sidebar on">
<button class="navbar-toggle"></button>
<nav class="navbar">
<div class="logo">
<a href="/"><img src="/images/ayer-side.svg" alt="Hy's Treasure house"></a>
</div>
<ul class="nav nav-main">
<li class="nav-item">
<a class="nav-item-link" href="/">主页</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/archives">归档</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/categories">分类</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/tags">标签</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/friends">友链</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/self_about">关于我</a>
</li>
<li class="nav-item">
<a class="nav-item-link" href="/blog_log">blog日志</a>
</li>
</ul>
</nav>
<nav class="navbar navbar-bottom">
<ul class="nav">
<li class="nav-item">
<a class="nav-item-link nav-item-search" title="搜索">
<i class="ri-search-line"></i>
</a>
</li>
</ul>
</nav>
<div class="search-form-wrap">
<div class="local-search local-search-plugin">
<input type="search" id="local-search-input" class="local-search-input" placeholder="Search...">
<div id="local-search-result" class="local-search-result"></div>
</div>
</div>
</aside>
<div id="mask"></div>
<!-- #reward -->
<div id="reward">
<span class="close"><i class="ri-close-line"></i></span>
<p class="reward-p"><i class="ri-cup-line"></i>请我喝杯咖啡吧~</p>
<div class="reward-box">
<div class="reward-item">
<img class="reward-img" src="/images/alipay.jpg">
<span class="reward-type">支付宝</span>
</div>
<div class="reward-item">
<img class="reward-img" src="/images/wechat.jpg">
<span class="reward-type">微信</span>
</div>
</div>
</div>
<script src="/js/jquery-3.6.0.min.js"></script>