-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathshowreel.html
2913 lines (1475 loc) · 140 KB
/
showreel.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
<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<!--
You have entered the realm of
▄▄▄▄ ██▀███ ▄▄▄ ███▄ █ ▓█████ ██████ ██░ ██ ▒█████ ██▓███
▓█████▄ ▓██ ▒ ██▒▒████▄ ██ ▀█ █ ▓█ ▀ ▒██ ▒ ▓██░ ██▒▒██▒ ██▒▓██░ ██▒
▒██▒ ▄██▓██ ░▄█ ▒▒██ ▀█▄ ▓██ ▀█ ██▒▒███ ░ ▓██▄ ▒██▀▀██░▒██░ ██▒▓██░ ██▓▒
▒██░█▀ ▒██▀▀█▄ ░██▄▄▄▄██ ▓██▒ ▐▌██▒▒▓█ ▄ ▒ ██▒░▓█ ░██ ▒██ ██░▒██▄█▓▒ ▒
░▓█ ▀█▓░██▓ ▒██▒ ▓█ ▓██▒▒██░ ▓██░░▒████▒▒██████▒▒░▓█▒░██▓░ ████▓▒░▒██▒ ░ ░
░▒▓███▀▒░ ▒▓ ░▒▓░ ▒▒ ▓▒█░░ ▒░ ▒ ▒ ░░ ▒░ ░▒ ▒▓▒ ▒ ░ ▒ ░░▒░▒░ ▒░▒░▒░ ▒▓▒░ ░ ░
▒░▒ ░ ░▒ ░ ▒░ ▒ ▒▒ ░░ ░░ ░ ▒░ ░ ░ ░░ ░▒ ░ ░ ▒ ░▒░ ░ ░ ▒ ▒░ ░▒ ░
░ ░ ░░ ░ ░ ▒ ░ ░ ░ ░ ░ ░ ░ ░ ░░ ░░ ░ ░ ▒ ░░
░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░
░
-->
<meta name="google-site-verification" content="cKie6SUiby1JmI2F8RMscJRBTa28kTM6XNHS5l1GXxc" />
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-131177872-1"></script>
<script type="text/javascript">
var host = window.location.hostname;
if ( host != "localhost") {
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-131177872-1');
}
</script>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
<title>Deep learning showreel! - Braneshop</title>
<meta name="geo.region" content="AU-VIC" />
<meta name="geo.placename" content="Melbourne" />
<link rel="stylesheet" type="text/css" href="./css/default.css?1" />
<link rel="stylesheet" type="text/css" href="./css/syntax.css" />
<link href="https://fonts.googleapis.com/css?family=PT+Mono|Lato|DM+Serif+Text|Karma|Gilda+Display|Quando" rel="stylesheet">
<script type="text/javascript" src="//cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
<!-- Meta tags for little cards and such. -->
<meta property="og:title" content="Deep learning showreel! - Braneshop" />
<meta property="og:url" content="https://braneshop.com.au/showreel.html" />
</head>
<body>
<div id="braneshop-fixed" onclick="document.location = '/';"></div>
<!-- Header -->
<div id="header">
<div id="header-top">
<h1><a href="./"><img id="header-logo" class="compressed" src="./images/frozen.png" title="Braneshop (Frozen)" alt="Braneshop (Frozen)" /></a></h1>
</div>
<h2>Expand your knowledge manifold!</h2>
</div>
<div class="menu-box">
<ul class="menu">
<li><a href="./" title="Home">Home</a></li>
<li><a href="./about.html" title="About">About</a></li>
<li><a href="./blog.html" title="Blog">Blog</a></li>
<li><a href="./showreel.html" title="Showreel">Showreel</a></li>
</ul>
</div>
<!-- Content -->
<div id="content">
<div class="section" id="showreel">
<h3> <span>Deep learning showreel! </span> </h3>
<div id="show-reel">
<p>This list is no-longer updated.</p>
<ul class="blog-list normal">
<li>
<div class="showreel-card">
<h5> <a href="./showreel/RigNet%20-%20Neural%20Rigging%20for%20Articulated%20Characters.html" rel="noopener">RigNet - Neural Rigging for Articulated Characters</a> — May 1, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/graphics.html">graphics</a>
| <small><a href="https://arxiv.org/pdf/2005.00559.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/RigNet - Neural Rigging for Articulated Characters.jpg"><img alt src="./images/showreel/RigNet - Neural Rigging for Articulated Characters.jpg" /></a></center>
<p>This is a truly awesome result (be sure to check out <a href="https://youtu.be/J90VETgWIDg">the video</a>) which can take in a 3D model, then compute a “rigging” so that it can be animated; i.e. walking, dancing, jumping, etc.</p>
<p>Super useful for animation and games, and just plain fun!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Investigating%20Transferability%20in%20Pretrained%20Language%20Models.html" rel="noopener">Investigating Transferability in Pretrained Language Models</a> — April 30, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/technical.html">technical</a>, <a href="./showreel-tags/text%20%28NLP%29.html">text (NLP)</a>
| <small><a href="https://arxiv.org/pdf/2004.14975.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Investigating Transferability in Pretrained Language Models.jpg"><img alt src="./images/showreel/Investigating Transferability in Pretrained Language Models.jpg" /></a></center>
<p>This is mostly a useful and interesting technical result around what is the best way to do transfer-learning when working with NLP models. The point is, in different data size regimes, you should do different things, to achieve the best results.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Consistent%20Video%20Depth%20Estimation.html" rel="noopener">Consistent Video Depth Estimation</a> — April 30, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/technical.html">technical</a>, <a href="./showreel-tags/video.html">video</a>
| <small><a href="https://arxiv.org/pdf/2004.15021.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Consistent Video Depth Estimation.jpg"><img alt src="./images/showreel/Consistent Video Depth Estimation.jpg" /></a></center>
<p>This is an interesting one. The idea is that we want to predict a “consistent” depth over the frames of a video. If we naively treat each frame as an independent image, we get very bad consistency; i.e. if we’re moving the camera across a scene; the depth changes wildly.</p>
<p>In the paper they introduce a technique to solve this problem, which then allows for very cool applications! Check out the <a href="https://roxanneluo.github.io/Consistent-Video-Depth-Estimation/">videos</a>.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Editing%20in%20Style%20-%20Uncovering%20the%20Local%20Semantics%20of%20GANs.html" rel="noopener">Editing in Style - Uncovering the Local Semantics of GANs</a> — April 29, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/technical.html">technical</a>
| <small><a href="https://arxiv.org/pdf/2004.14367.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Editing in Style - Uncovering the Local Semantics of GANs.jpg"><img alt src="./images/showreel/Editing in Style - Uncovering the Local Semantics of GANs.jpg" /></a></center>
<p>Generative adversarial networks (GANs) are hugely popular for generating fake images. But a key topic is fine-grained editing and adjusting. I.e. you can generate a random person, but can you generate a random <em>smiling</em> person? To what degree can we edit these visual features of people?</p>
<p>This paper presents a mechanism by using reference images and regions and transferring them onto the generated images. Neat! And useful.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/MakeItTalk%20-%20Speaker-Aware%20Talking%20Head%20Animation.html" rel="noopener">MakeItTalk - Speaker-Aware Talking Head Animation</a> — April 27, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/audio.html">audio</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/video.html">video</a>
| <small><a href="https://arxiv.org/pdf/2004.12992.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/MakeItTalk - Speaker-Aware Talking Head Animation.jpg"><img alt src="./images/showreel/MakeItTalk - Speaker-Aware Talking Head Animation.jpg" /></a></center>
<p>A simple dream realised. We take an audoclip, and a still image, and animate the image as if it’s speaking the audio. Check out <a href="https://youtu.be/OU6Ctzhpc6s">the video</a>.</p>
<p>Adobe is investing heavily in this kind of connection between audio and images; they also did some early work on “audio inpainting”. Expect some cool products from them in the future.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Graph2Plan%20-%20Learning%20Floorplan%20Generation%20from%20Layout%20Graphs.html" rel="noopener">Graph2Plan - Learning Floorplan Generation from Layout Graphs</a> — April 27, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/architecture.html">architecture</a>, <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/graphs%20%28GNN%29.html">graphs (GNN)</a>
| <small><a href="https://arxiv.org/pdf/2004.13204.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Graph2Plan - Learning Floorplan Generation from Layout Graphs.jpg"><img alt src="./images/showreel/Graph2Plan - Learning Floorplan Generation from Layout Graphs.jpg" /></a></center>
<p>I have a soft-spot for AI-generated architecture; and this one is interesting because it gives the user some high-level control on what the generation does. I.e. you can control the number of rooms, how they connect, and roughly the layout.</p>
<p>This kind of thing, built well into modern CAD tools, would be really cool!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Quantum%20Gradient%20Algorithm%20for%20General%20Polynomials.html" rel="noopener">Quantum Gradient Algorithm for General Polynomials</a> — April 23, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/quantum.html">quantum</a>
| <small><a href="https://arxiv.org/pdf/2004.11086.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Quantum Gradient Algorithm for General Polynomials.jpg"><img alt src="./images/showreel/Quantum Gradient Algorithm for General Polynomials.jpg" /></a></center>
<p>The idea here is that, in principle, if you have a lot of parameters in your function, then the time to compute the entire gradient w.r.t. all the parameters will scale like the number of parameters. This is bad when you have billions of parameters; so mostly in optimisation we focus on <em>stochastic</em> gradient descent; i.e. just looking at a small number of parameters at any one time.</p>
<p>Here, the idea is that we could use quantum computers to do the full computation significantly faster. In this, and algorithm is introduced that in fact achieves this, for general polynomials (perhaps a step towards achieving it for full neural networks).</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/ViBE%20-%20A%20Tool%20for%20Measuring%20and%20Mitigating%20Bias%20in%20Image%20Datasets.html" rel="noopener">ViBE - A Tool for Measuring and Mitigating Bias in Image Datasets</a> — April 16, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/bias.html">bias</a>, <a href="./showreel-tags/ethics.html">ethics</a>
| <small><a href="https://arxiv.org/pdf/2004.07999.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/ViBE - A Tool for Measuring and Mitigating Bias in Image Datasets.jpg"><img alt src="./images/showreel/ViBE - A Tool for Measuring and Mitigating Bias in Image Datasets.jpg" /></a></center>
<p>This is an <a href="https://github.com/princetonvisualai/vibe-tool">interesting tool</a> that helps formalise how to think about bias in your datasets. It allows you to gain formalised insights (i.e. “you don’t have enough data from this country”) and then generate concrete actions. It’s early days, but tools like this will be very useful as the field matures.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Efficient%20State%20Read-out%20for%20Quantum%20Machine%20Learning%20Algorithms.html" rel="noopener">Efficient State Read-out for Quantum Machine Learning Algorithms</a> — April 14, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/quantum.html">quantum</a>
| <small><a href="https://arxiv.org/pdf/2004.06421.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Efficient State Read-out for Quantum Machine Learning Algorithms.jpg"><img alt src="./images/showreel/Efficient State Read-out for Quantum Machine Learning Algorithms.jpg" /></a></center>
<p>One annoying technical issue that occurs in quantum machine learning is that, <em>maybe</em> it’s possible to run a quantum algorithm that performs better than a classical one, but you <em>lose</em> all the speedup when you simply try and <em>read</em> the answer out! So frustrating.</p>
<p>This paper addresses this problem.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/3D%20Photography%20using%20Context-aware%20Layered%20Depth%20Inpainting.html" rel="noopener">3D Photography using Context-aware Layered Depth Inpainting</a> — April 9, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>
| <small><a href="https://arxiv.org/pdf/2004.04727.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/3D Photography using Context-aware Layered Depth Inpainting.jpg"><img alt src="./images/showreel/3D Photography using Context-aware Layered Depth Inpainting.jpg" /></a></center>
<p>A nice technique to make a cool little animation from a 2D image. <a href="https://shihmengli.github.io/3D-Photo-Inpainting/">Check out the video here</a>.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Typilus%3A%20Neural%20Type%20Hints.html" rel="noopener">Typilus: Neural Type Hints</a> — April 6, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/code-understanding.html">code-understanding</a>, <a href="./showreel-tags/graphs%20%28GNN%29.html">graphs (GNN)</a>, <a href="./showreel-tags/text%20%28NLP%29.html">text (NLP)</a>
| <small><a href="https://arxiv.org/pdf/2004.10657.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Typilus: Neural Type Hints.jpg"><img alt src="./images/showreel/Typilus: Neural Type Hints.jpg" /></a></center>
<p>This is probably a niche one; but interesting to see someone tackling this problem. The idea is that estimating types for Python functions is hard; so perhaps a neural network can be used to do it. Turns out, you can go some way with this idea!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Learning%20Generative%20Models%20of%20Shape%20Handles.html" rel="noopener">Learning Generative Models of Shape Handles</a> — April 6, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="https://arxiv.org/pdf/2004.03028.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Learning Generative Models of Shape Handles.jpg"><img alt src="./images/showreel/Learning Generative Models of Shape Handles.jpg" /></a></center>
<p>You can imagine that manipulating, arbitrarily, objects in 3D is very hard. How to change a chair in a “structured” way? This paper presents a really cool way of manipulating such shapes; with pretty cool results!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/GANSpace%20-%20Discovering%20Interpretable%20GAN%20Controls.html" rel="noopener">GANSpace - Discovering Interpretable GAN Controls</a> — April 6, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/technical.html">technical</a>
| <small><a href="https://arxiv.org/pdf/2004.02546.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/GANSpace - Discovering Interpretable GAN Controls.jpg"><img alt src="./images/showreel/GANSpace - Discovering Interpretable GAN Controls.jpg" /></a></center>
<p>A theme that comes up again and again; this is another great piece of work in this vein; how can we control the generated images from GANs?</p>
<p>This is really a very important topic, in part because it will empower users to interact with these tools.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Adaptive%20Fractional%20Dilated%20Convolution%20Network%20for%20Image%20Aesthetics%20Assessment.html" rel="noopener">Adaptive Fractional Dilated Convolution Network for Image Aesthetics Assessment</a> — April 6, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/technical.html">technical</a>
| <small><a href="https://arxiv.org/pdf/2004.03015.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Adaptive Fractional Dilated Convolution Network for Image Aesthetics Assessment.jpg"><img alt src="./images/showreel/Adaptive Fractional Dilated Convolution Network for Image Aesthetics Assessment.jpg" /></a></center>
<p>I like this idea because it reminds me of Christopher Alexander’s ideas of the objective beauty of photos and the idea of considering which photo has more “life”.</p>
<p>It’s also useful for judging the quality of generative images; which is the most asthetic?</p>
<p>Interesting work.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Generating%20Rationales%20in%20Visual%20Question%20Answering.html" rel="noopener">Generating Rationales in Visual Question Answering</a> — April 4, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/explainability%20%28XAI%29.html">explainability (XAI)</a>, <a href="./showreel-tags/text%20%28NLP%29.html">text (NLP)</a>, <a href="./showreel-tags/visual-question-answering%20%28VQA%29.html">visual-question-answering (VQA)</a>
| <small><a href="https://arxiv.org/pdf/2004.02032.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Generating Rationales in Visual Question Answering.jpg"><img alt src="./images/showreel/Generating Rationales in Visual Question Answering.jpg" /></a></center>
<p>A nice result in the field of explainability; we can give visual explanations to visual questions, to help explain our reasoning. Very useful.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Learning%20to%20See%20Through%20Obstructions.html" rel="noopener">Learning to See Through Obstructions</a> — April 2, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>
| <small><a href="https://arxiv.org/pdf/2004.01180.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Learning to See Through Obstructions.jpg"><img alt src="./images/showreel/Learning to See Through Obstructions.jpg" /></a></center>
<p>Very simple one; a computer-vision algorithm to just remove general visual obstructions (like rain, glare, fences) from photos! Handy.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Sign%20Language%20Translation%20With%20Transformers.html" rel="noopener">Sign Language Translation With Transformers</a> — April 1, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/real-world.html">real-world</a>
| <small><a href="https://arxiv.org/pdf/2004.00588.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Sign Language Translation With Transformers.jpg"><img alt src="./images/showreel/Sign Language Translation With Transformers.jpg" /></a></center>
<p>Nice technique to watch sign language videos and interpreting them.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Design%20of%20Variational%20Autoencoder-based%20Rhythm%20Generator%20as%20a%20DAW%20plugin.html" rel="noopener">Design of Variational Autoencoder-based Rhythm Generator as a DAW plugin</a> — April 1, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/music.html">music</a>
| <small><a href="https://arxiv.org/pdf/2004.01525.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Design of Variational Autoencoder-based Rhythm Generator as a DAW plugin.jpg"><img alt src="./images/showreel/Design of Variational Autoencoder-based Rhythm Generator as a DAW plugin.jpg" /></a></center>
<p>I like this because it’s a kind of “applied” example of generative AI; here they build a generative music AI into Abelton Live. <a href="https://magenta.tensorflow.org/studio/">Check out the code here</a>.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Convolutional%20Neural%20Networks%20for%20Image-based%20Corn%20Kernel%20Detection%20and%20Counting.html" rel="noopener">Convolutional Neural Networks for Image-based Corn Kernel Detection and Counting</a> — March 26, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/real-world.html">real-world</a>
| <small><a href="https://arxiv.org/pdf/2003.12025.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Convolutional Neural Networks for Image-based Corn Kernel Detection and Counting.jpg"><img alt src="./images/showreel/Convolutional Neural Networks for Image-based Corn Kernel Detection and Counting.jpg" /></a></center>
<p>I like this because it’s just a simple idea, implemented well, that is probably of use to someone!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Deformable%20Style%20Transfer.html" rel="noopener">Deformable Style Transfer</a> — March 24, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="https://arxiv.org/pdf/2003.11038.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Deformable Style Transfer.jpg"><img alt src="./images/showreel/Deformable Style Transfer.jpg" /></a></center>
<p>I really think this is quite amazing. The results of the style transfer on these phtoso are simply the best I’ve ever seen. Incredible!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Learning%20Object%20Permanence%20from%20Video.html" rel="noopener">Learning Object Permanence from Video</a> — March 23, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/technical.html">technical</a>
| <small><a href="https://arxiv.org/pdf/2003.10469.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Learning Object Permanence from Video.jpg"><img alt src="./images/showreel/Learning Object Permanence from Video.jpg" /></a></center>
<p>A classic thing that babies learn is that objects don’t physically disappear when they go behind things. Neural networks have struggled with this problem. This paper goes some way to addressing this issue.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/3D-CariGAN%20-%20An%20End-to-End%20Solution%20to%203D%20Caricature%20Generation%20from%20Face%20Photos.html" rel="noopener">3D-CariGAN - An End-to-End Solution to 3D Caricature Generation from Face Photos</a> — March 15, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/real-world.html">real-world</a>
| <small><a href="https://arxiv.org/pdf/2003.06841.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/3D-CariGAN - An End-to-End Solution to 3D Caricature Generation from Face Photos.jpg"><img alt src="./images/showreel/3D-CariGAN - An End-to-End Solution to 3D Caricature Generation from Face Photos.jpg" /></a></center>
<p>I think this one is just plain funny. An end-to-end approach for generating a 3d physical caricature from a photo. Cool!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Towards%20Photo-Realistic%20Virtual%20Try-On%20by%20Adaptively%20Generating%20and%20Preserving%20Image%20Content.html" rel="noopener">Towards Photo-Realistic Virtual Try-On by Adaptively Generating and Preserving Image Content</a> — March 12, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/fashion.html">fashion</a>
| <small><a href="https://arxiv.org/pdf/2003.05863.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Towards Photo-Realistic Virtual Try-On by Adaptively Generating and Preserving Image Content.jpg"><img alt src="./images/showreel/Towards Photo-Realistic Virtual Try-On by Adaptively Generating and Preserving Image Content.jpg" /></a></center>
<p>This is a nice step towards high-quality “virtual try-on”; i.e. the idea that you can take an image of a person, and a product photo, and then see what it would look like to “try” that item on!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/TensorFlow%20Quantum.html" rel="noopener">TensorFlow Quantum</a> — March 6, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/quantum.html">quantum</a>, <a href="./showreel-tags/tool.html">tool</a>
| <small><a href="https://arxiv.org/pdf/2003.02989.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/TensorFlow Quantum.jpg"><img alt src="./images/showreel/TensorFlow Quantum.jpg" /></a></center>
<p>This has been coming for a while; it’s a start of building some kind of quantum capability into the TensorFlow library. It’s not perfect, but it’s a start.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Privacy-preserving%20Learning%20via%20Deep%20Net%20Pruning.html" rel="noopener">Privacy-preserving Learning via Deep Net Pruning</a> — March 4, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/privacy.html">privacy</a>, <a href="./showreel-tags/technical.html">technical</a>
| <small><a href="https://arxiv.org/pdf/2003.01876.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Privacy-preserving Learning via Deep Net Pruning.jpg"><img alt src="./images/showreel/Privacy-preserving Learning via Deep Net Pruning.jpg" /></a></center>
<p>This is interesting paper in that it shows how “differential privacy” can be related to network pruning. Differential privacy is the idea that we can hide individual datapoints by adding noise. This is useful if, say, working on medical data where we want to not reveal individual patients.</p>
<p>The trade-off is between how much worse does the network get as we increase the privacy (i.e. make it harder to recover individual datapoints). This paper links this idea with the idea of pruning neural networks in a precise way.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Learning%20to%20Transfer%20Texture%20from%20Clothing%20Images%20to%203D%20Humans.html" rel="noopener">Learning to Transfer Texture from Clothing Images to 3D Humans</a> — March 4, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/fashion.html">fashion</a>
| <small><a href="https://arxiv.org/pdf/2003.02050.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Learning to Transfer Texture from Clothing Images to 3D Humans.jpg"><img alt src="./images/showreel/Learning to Transfer Texture from Clothing Images to 3D Humans.jpg" /></a></center>
<p>A good step on the way to <em>full</em> virtual try-on would be just to dress 3D models with product photos. This paper does just that!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Inverse%20Graphics%20GAN%20-%20Learning%20to%20Generate%203D%20Shapes%20from%20Unstructured%202D%20Data.html" rel="noopener">Inverse Graphics GAN - Learning to Generate 3D Shapes from Unstructured 2D Data</a> — February 28, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="http://arxiv.org/pdf/2002.12674.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Inverse Graphics GAN - Learning to Generate 3D Shapes from Unstructured 2D Data.jpg"><img alt src="./images/showreel/Inverse Graphics GAN - Learning to Generate 3D Shapes from Unstructured 2D Data.jpg" /></a></center>
<p>This is a cool idea done well. The point is, you can imagine it’s very hard to generate an arbitrary 3D model for a given image. This does a pretty good job!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Sketch-to-Art%20-%20Synthesizing%20Stylized%20Art%20Images%20From%20Sketches%20.html" rel="noopener">Sketch-to-Art - Synthesizing Stylized Art Images From Sketches</a> — February 26, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/art.html">art</a>, <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="https://arxiv.org/pdf/2002.12888.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Sketch-to-Art - Synthesizing Stylized Art Images From Sketches .jpg"><img alt src="./images/showreel/Sketch-to-Art - Synthesizing Stylized Art Images From Sketches .jpg" /></a></center>
<p>This is just a nice result. I really like the outputs that are generated. You can play around with this idea <a href="https://create.playform.io/sketch-to-image">over here</a>.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Wait%2C%20Im%20Still%20Talking.html" rel="noopener">Wait, I'm Still Talking</a> — February 22, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/text%20%28NLP%29.html">text (NLP)</a>
| <small><a href="https://arxiv.org/pdf/2002.09616.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Wait, I'm Still Talking.jpg"><img alt src="./images/showreel/Wait, I'm Still Talking.jpg" /></a></center>
<p>This is a nice result that estimates the best time to respond to a series of chat messages. In a idealised world maybe every chat message is atomic; but in a real chat situation, people can ramble over several messages until they get to their central question.</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/AutoFoley%20-%20Artificial%20Synthesis%20of%20Synchronized%20Sound%20Tracks%20for%20Silent%20Videos%20with%20Deep%20Learning.html" rel="noopener">AutoFoley - Artificial Synthesis of Synchronized Sound Tracks for Silent Videos with Deep Learning</a> — February 21, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/audio.html">audio</a>, <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/real-world.html">real-world</a>, <a href="./showreel-tags/video.html">video</a>
| <small><a href="https://arxiv.org/pdf/2002.10981.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/AutoFoley - Artificial Synthesis of Synchronized Sound Tracks for Silent Videos with Deep Learning.jpg"><img alt src="./images/showreel/AutoFoley - Artificial Synthesis of Synchronized Sound Tracks for Silent Videos with Deep Learning.jpg" /></a></center>
<p>This is a cool idea; given some video, generate the sound. The paper is worth a glance for the photo they show of a foley studio! I never knew they looked so cool!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Seeing%20the%20World%20in%20a%20Bag%20of%20Chips.html" rel="noopener">Seeing the World in a Bag of Chips</a> — January 14, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>
| <small><a href="https://arxiv.org/pdf/2001.04642.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Seeing the World in a Bag of Chips.jpg"><img alt src="./images/showreel/Seeing the World in a Bag of Chips.jpg" /></a></center>
<p>Is the world reflected in a bag of chips? Yes! Turns out, using a neural network, if you want to know whats going on around you; you can just take a photo (with depth information) of the chip packet and figure it out!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Learning%20to%20Zoom-in%20via%20Learning%20to%20Zoom-out.html" rel="noopener">Learning to Zoom-in via Learning to Zoom-out</a> — January 8, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>
| <small><a href="https://arxiv.org/pdf/2001.02381.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Learning to Zoom-in via Learning to Zoom-out.jpg"><img alt src="./images/showreel/Learning to Zoom-in via Learning to Zoom-out.jpg" /></a></center>
<p>This paper tackles the now-standard problem of “enhancing” an image; i.e. being able to look into any region of an image in greater detail. The trick here is that they utilise the idea of learning to zoom <em>out</em> to also learn to zoom <em>in</em>. The results are very good!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Painting%20Many%20Pasts%20-%20Synthesizing%20Time%20Lapse%20Videos%20of%20Paintings.html" rel="noopener">Painting Many Pasts - Synthesizing Time Lapse Videos of Paintings</a> — January 4, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/art.html">art</a>, <a href="./showreel-tags/generative.html">generative</a>, <a href="./showreel-tags/video.html">video</a>
| <small><a href="https://arxiv.org/pdf/2001.01026.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Painting Many Pasts - Synthesizing Time Lapse Videos of Paintings.jpg"><img alt src="./images/showreel/Painting Many Pasts - Synthesizing Time Lapse Videos of Paintings.jpg" /></a></center>
<p>This is a cute one. The idea is simply to see if it’s’ possible to show the “painting progression” of a finished painting. I.e. what might be (photographs) of the steps of making this painting?</p>
<p>I think this is a really neat idea; would love to see some artists try and replicate the steps, to confirm that they (somewhat) match reality!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/Generating%20Object%20Stamps.html" rel="noopener">Generating Object Stamps</a> — January 1, 2020 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="https://arxiv.org/pdf/2001.02595.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/Generating Object Stamps.jpg"><img alt src="./images/showreel/Generating Object Stamps.jpg" /></a></center>
<p>Here we have the idea of, given some background image, can we fill it with arbitrary objects so that it looks realistic? Here they show that they are able to solve this problem very nicely, by filling a diverse range of wildlife scenes with animals!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/FaceShifter%20-%20Towards%20High%20Fidelity%20And%20Occlusion%20Aware%20Face%20Swapping.html" rel="noopener">FaceShifter - Towards High Fidelity And Occlusion Aware Face Swapping</a> — December 31, 2019 </h5>
<span class="tags">Tags: <a href="./showreel-tags/computer-vision.html">computer-vision</a>, <a href="./showreel-tags/generative.html">generative</a>
| <small><a href="https://arxiv.org/pdf/1912.13457.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/FaceShifter - Towards High Fidelity And Occlusion Aware Face Swapping.jpg"><img alt src="./images/showreel/FaceShifter - Towards High Fidelity And Occlusion Aware Face Swapping.jpg" /></a></center>
<p>I think the results of this one are amazing. They get this to work using a multi-staged network approach; i.e. a few different networks performing different functions, such as one to learn “attributes”, and another to refine the output of the main one. Lots to learn from this!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/SketchTransfer%20-%20A%20Challenging%20New%20Task%20for%20Exploring%20Detail-Invariance%20and%20the%20Abstractions%20Learned%20by%20Deep%20Networks.html" rel="noopener">SketchTransfer - A Challenging New Task for Exploring Detail-Invariance and the Abstractions Learned by Deep Networks</a> — December 25, 2019 </h5>
<span class="tags">Tags: <a href="./showreel-tags/dataset.html">dataset</a>
| <small><a href="https://arxiv.org/pdf/1912.11570.pdf">Paper</a></small>
</span>
<div class="content">
<center><a href="./images/showreel/SketchTransfer - A Challenging New Task for Exploring Detail-Invariance and the Abstractions Learned by Deep Networks.jpg"><img alt src="./images/showreel/SketchTransfer - A Challenging New Task for Exploring Detail-Invariance and the Abstractions Learned by Deep Networks.jpg" /></a></center>
<p>This isn’t a network itself; it’s a dataset and a proposed task. The idea of this dataset is that there is definitely a similarity between <em>sketches</em> of real objects, and photos of those objects; but, neural networks aren’t great at knowing this abstraction, yet. I.e. If I have an object-classifier, it can’t very well classify sketches. We’ll see how people progress towards this problem via this dataset!</p>
</div>
</div>
</li>
<li>
<div class="showreel-card">
<h5> <a href="./showreel/ConvPoseCNN%20-%20Dense%20Convolutional%206D%20Object%20Pose%20Estimation.html" rel="noopener">ConvPoseCNN - Dense Convolutional 6D Object Pose Estimation</a> — December 16, 2019 </h5>
<span class="tags">Tags: <a href="./showreel-tags/3d.html">3d</a>, <a href="./showreel-tags/pose.html">pose</a>
| <small><a href="https://arxiv.org/pdf/1912.07333.pdf">Paper</a></small>
</span>
<div class="content">