-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
1102 lines (1024 loc) · 76.1 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="en"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Rameswar Panda</title>
<meta name="author" content="Rameswar Panda">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
</head>
<body>
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:0px">
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:2.5%;width:63%;vertical-align:middle">
<p style="text-align:center">
<name>Rameswar Panda</name>
</p>
<p>I am a research scientist at <a href="https://mitibmwatsonailab.mit.edu/">MIT-IBM Watson AI Lab</a>, where I work on computer vision and machine learning.
</p>
<p>
I received my Ph.D. from <a href="http://www.ucr.edu">UC Riverside</a> in 2018, under the supervision of <a href="http://vcg.engr.ucr.edu/amit">Prof. Amit K. Roy-Chowdhury</a>.
<!-- Previously, I obtained my M.S. from <a href="http://www.jaduniv.edu.in/">Jadavpur University</a> in 2013 supervised by <a href="https://sites.google.com/site/anandachowdhury/">Prof. Ananda S. Chowdhury</a>.-->
During Ph.D., I was very fortunate to have interned at <a href="http://www.nec-labs.com/">NEC Labs</a>, <a href="https://research.adobe.com/">Adobe Research</a> and <a href="http://www.usa.siemens.com/en/about_us/research/home.htm">Siemens Research</a>.
</p>
<p style="text-align:center">
<a href="mailto:[email protected]">Email</a>  / 
<a href="data/CV_Rameswar.pdf">CV</a>  / 
<a href="https://scholar.google.co.in/citations?user=_ySuu6gAAAAJ&hl=en">Google Scholar</a>  / 
<a href="https://www.linkedin.com/in/rameswar-panda-0067b2ab">LinkedIn</a>
</p>
<!-- <br>-->
<!-- <p>-->
<!-- <span style="color: #ff0000; "><strong>[Internship]</strong></span> We are currently looking to hire multiple PhD candidates for 2023 summer internship, to conduct fundamental research in many areas of computer vision, machine learning-->
<!-- and natural language processing. Please send me an <a href="mailto:[email protected]">email</a> if you are interested.-->
<!-- </p>-->
</td>
<td style="padding:2.5%;width:40%;max-width:40%">
<a href="images/myphoto_circle.png"><img style="width:90%;max-width:90%" alt="profile photo" src="images/myphoto_circle.png" class="hoverZoomLink"></a>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Research</heading>
<p>
My research interests mainly lie in the areas of computer vision, machine learning and more recently natural language processing, with a special focus on
efficient AI, including data efficiency and model efficiency.
<!-- In particular, my current focus is on making AI systems more efficient, i.e., developing novel deep learning methods that can operate with less human-annotated data (data efficient), and less computation (model efficient).-->
I am also interested in image/video understanding, unsupervised/self-supervised representation learning and multimodal learning.
</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tr>
<td style="padding:2.5%;width:100%;vertical-align:middle">
<heading>News</heading>
<p></p>
<div class="listbox">
<ul>
<li> Three papers accepted at <a href="https://iclr.cc/Conferences/2023">ICLR 2023</a>. One of them as "notable-top-25%" paper (Spotlight).
<li> Select, Label, and Mix (SLM) received the Best Paper Honorable Mention Award at <a href="https://wacv2023.thecvf.com/">WACV 2023</a>.
<li> Two papers accepted at <a href="https://wacv2023.thecvf.com/">WACV 2023</a>.
<li> Two papers accepted at <a href="https://nips.cc/">NeurIPS 2022</a>.
<li> One paper accepted at <a href="https://icml.cc/">ICML 2022</a>.
<li> Two papers accepted at <a href="http://cvpr2022.thecvf.com/">CVPR 2022</a>.
<li> Two papers accepted at <a href="https://iclr.cc/Conferences/2022">ICLR 2022</a>.
<li> I am serving as an Area Chair for <a href="http://wacv2022.thecvf.com/home">WACV 2022</a> and <a href="https://events.iitgn.ac.in/2022/icvgip/">ICVGIP 2022</a>.<br>
<li> We are organizing the 2nd workshop on <a href="https://sites.google.com/view/cvpr2022-dnetcv/">Dynamic Neural Networks</a> at <a href="http://cvpr2022.thecvf.com/">CVPR 2022</a>.<br>
<li> I am serving as a Sponsorship Chair for <a href="http://vcip2022.org/index.html">VCIP 2022</a>.<br>
<li> Three papers accepted at <a href="https://neurips.cc/">NeurIPS 2021</a>.
<li> Six papers accepted at <a href="http://iccv2021.thecvf.com/">ICCV 2021</a>. Two of them are selected for oral presentation.<br>
<li> We are organizing a tutorial on <a href="https://sites.google.com/view/effvideo-2021/">Efficient Video Understanding</a> at <a href="http://iccv2021.thecvf.com/">ICCV 2021</a>.<br>
<li> Paper on Fair Selective Classification accepted at <a href="https://icml.cc/">ICML 2021</a>.<br>
<li> Two papers on Video Action Recognition accepted at <a href="http://cvpr2021.thecvf.com/">CVPR 2021</a>.<br>
<li> Two papers on Efficient Video Understanding accepted at <a href="https://iclr.cc/Conferences/2021">ICLR 2021</a>.<br>
<li> We are organizing a workshop on <a href="https://sites.google.com/view/cvpr2021-dnetcv/">Dynamic Neural Networks Meets Computer Vision (DNetCV)</a> at <a href="http://cvpr2021.thecvf.com/">CVPR 2021</a>.<br>
<li> We are organizing the 2nd workshop and challenges on <a href="http://cvpr21-nas.com/">Neural Architecture Search (NAS)</a> at <a href="http://cvpr2021.thecvf.com/">CVPR 2021</a>.<br>
<li> Paper on Analyzing Transferability in Neural Architecture Search accepted at <a href="https://aaai.org/Conferences/AAAI-21/">AAAI 2021</a>.<br>
<li> Paper on Learning What to Share for Efficient Deep Multi-Task Learning accepted at <a href="https://nips.cc/">NeurIPS 2020</a>.<br>
<li> Paper on Unsupervised Video Person Re-Identification accepted at <a href="https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=76">IEEE TCSVT, 2020</a>.<br><li> Paper on Adversarial Knowledge Transfer from Unlabeled Data accepted at <a href="https://2020.acmmm.org/">ACM MM 2020</a>.</li>
<li> Paper on Adaptive Frame Resolution for Efficient Action Recognition accepted at <a href="https://eccv2020.eu/">ECCV 2020</a>.<br>
<li> Paper on Fairness of Classifiers Across Skin Tones in Dermatology accepted at <a href="https://www.miccai2020.org/en/">MICCAI 2020</a>.<br>
<li> We are organizing the 2nd edition of our <a href="https://sites.google.com/view/multimodalvideo-v2/">Workshop on Multi-Modal Video Analysis</a> at <a href="https://eccv2020.eu/">ECCV 2020</a>.<br>
<li> Paper on Non-Adversarial Video Synthesis with Learned Priors accepted at <a href="http://cvpr2020.thecvf.com/">CVPR 2020</a>.<br>
<li> Paper on Hypothesis Transfer Learning for Person Re-ID accepted at <a href="http://cvpr2020.thecvf.com/">CVPR 2020</a>.<br>
</ul>
</div>
</td>
</tr>
</table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<heading>Publications</heading>
</td>
</tr>
</tbody></table>
<!-- <div style="width: 100%; height: 20px; border-bottom: 1px solid black; text-align: center">-->
<!-- <span style="font-size: 20px; background-color: #F3F5F6; padding: 0 10px;">-->
<!-- 2020 <!–Padding is optional–>-->
<!-- </span>-->
<!-- </div> <br>-->
<!-- <table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>-->
<!-- <tr>-->
<!-- <td style="padding:20px;width:100%;vertical-align:middle">-->
<!-- <!– <h2><b><font color="red">2021</font></b></h2> –>-->
<!-- <h2>2022</h2>-->
<!-- </td>-->
<!-- </tr>-->
<!-- </tbody></table>-->
<!-- Horizontal space:     -->
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/WACV_2023_slm.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/WACV_2023_slm.pdf>
<papertitle>Select, Label, and Mix: Learning Discriminative Invariant Feature Representations for Partial Domain Adaptation</papertitle>
</a>
<br>
Aadarsh Sahoo, <strong>Rameswar Panda</strong>, Rogerio Feris, Kate Saenko, Abir Das<br>
<em>Winter Conference on Applications of Computer Vision (<strong>WACV</strong>)</em>, 2023<br>
<em>NeurIPS Workshop on Distribution Shifts (<strong>NeurIPS-W</strong>)</em>, 2021 [<a href="https://rpand002.github.io/data/NeurIPS_2021_W_SLM.pdf">PDF</a>]<br>
<span style="color: #ff0000; "><strong>Best Paper Honorable Mention</strong></span> <br>
[<a href="https://cvir.github.io/projects/slm.html">Project Page</a>] [<a href="https://github.com/CVIR/SLM">Code</a>] [<a href="https://rpand002.github.io/data/WACV_2023_slm_supp.pdf">Supplementary Material</a>]<br>
<p>We develop a new framework for learning discriminative and invariant feature representation while preventing intrinsic negative transfer in partial domain adaptation.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/WACV_2023_simultaneous.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/WACV_2023_simultaneous.pdf>
<papertitle>Semi-Supervised Domain Adaptation with Auto-Encoder via Simultaneous Learning</papertitle>
</a>
<br>
Md Mahmudur Rahman, <strong>Rameswar Panda</strong>, Mohammad Arif Ul Alam<br>
<em>Winter Conference on Applications of Computer Vision (<strong>WACV</strong>)</em>, 2023<br>
<p> We propose an auto-encoder-based approach combined with a simultaneous learning scheme to align source and target features for semi-supervised domain adaptation.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/NeurIPS_2022_synapt.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2022_synapt.pdf>
<papertitle>How Transferable are Video Representations Based on Synthetic Data?</papertitle>
</a>
<br>
Yo-whan Kim, Samarth Mishra, SouYoung Jin, <strong>Rameswar Panda</strong>, Hilde Kuehne, Leonid Karlinsky, Venkatesh Saligrama, Kate Saenko, Aude Oliva, Rogerio Feris<br>
<em>Neural Information Processing Systems Datasets (<strong>NeurIPS</strong>)</em>, 2022<br>
[<a href="https://github.com/mintjohnkim/SynAPT">Dataset</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2022_synapt_supp.pdf">Supplementary Material</a>]<br>
<p> We propose a new benchmark, SynAPT, for studying the transferability of synthetic video representations for action recognition.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/NeurIPS_2022_feta.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2022_feta.pdf>
<papertitle>FETA: Towards Specializing Foundational Models for Expert Task Applications</papertitle>
</a>
<br>
Amit Alfassy, Assaf Arbelle, Oshri Halimi, Sivan Harary, Roei Herzig, Eli Schwartz, <strong>Rameswar Panda</strong>, Michele Dolfi, Christoph Auer, Peter Staar, Kate Saenko, Rogerio Feris, Leonid Karlinsky<br>
<em>Neural Information Processing Systems Datasets (<strong>NeurIPS</strong>)</em>, 2022<br>
[<a href="https://ai-vision-public-datasets.s3.eu.cloud-object-storage.appdomain.cloud/FETA/feta.tar.gz">Dataset</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2022_feta_supp.pdf">Supplementary Material</a>]<br>
<p>We introduce FETA benchmark to understand technical documentations, via learning to match their graphical illustrations to corresponding language descriptions.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICML_2022_selective.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICML_2022_selective.pdf>
<papertitle>Selective Regression Under Fairness Criteria</papertitle>
</a>
<br>
Abhin Shah, Yuheng Bu, Joshua K Lee, Subhro Das, <strong>Rameswar Panda</strong>, Prasanna Sattigeri, Gregory Wornell<br>
<em>International Conference on Machine Learning (<strong>ICML</strong>)</em>, 2022<br>
<p> We demonstrate the performance disparities across different subgroups for selective regression and develop novel methods to mitigate such disparities.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2022_valhalla.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2022_valhalla.pdf>
<papertitle>VALHALLA: Visual Hallucination for Machine Translation</papertitle>
</a>
<br>
Yi Li, <strong>Rameswar Panda</strong>, Yoon Kim, Chun-Fu (Richard) Chen, Rogerio Feris, David Cox, Nuno Vasconcelos<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2022<br>
[<a href="http://www.svcl.ucsd.edu/projects/valhalla">Project Page</a>] [<a href="https://github.com/JerryYLi/valhalla-nmt">Code</a>] [<a href="https://rpand002.github.io/data/CVPR_2022_valhalla_supp.pdf">Supplementary Material</a>]<br>
<p> We introduce a visual hallucination framework which requires only source sentences at inference time and instead uses hallucinated visual representations for multimodal machine translation.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2022_task2sim.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2022_task2sim.pdf>
<papertitle>Task2Sim: Towards Effective Pre-training and Transfer from Synthetic Data</papertitle>
</a>
<br>
Samarth Mishra, <strong>Rameswar Panda</strong>, Cheng Perng Phoo, Chun-Fu (Richard) Chen, Leonid Karlinsky, Kate Saenko, Venkatesh Saligrama, Rogerio Feris<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2022<br>
[<a href="https://samarth4149.github.io/projects/task2sim.html">Project Page</a>] [<a href="https://github.com/samarth4149/task2sim">Code</a>] [<a href="https://rpand002.github.io/data/CVPR_2022_task2sim_supp.pdf">Supplementary Material</a>]<br>
<p> We propose a unified model mapping downstream task representations to optimal simulation parameters to generate synthetic pre-training data for them.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICLR_2022_regionvit.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICLR_2022_regionvit.pdf>
<papertitle>RegionViT: Regional-to-Local Attention for Vision Transformers</papertitle>
</a>
<br>
Chun-Fu (Richard) Chen, <strong>Rameswar Panda</strong>, Quanfu Fan<br>
<em> International Conference on Learning Representations (<strong>ICLR</strong>)</em>, 2022 <br>
[<a href="https://github.com/ibm/regionvit">Code</a>]<br>
<p> We propose a new architecture that adopts the pyramid structure and employ a regional-to-local attention rather than global self-attention in vision transformers.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICLR_2022_sifar.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICLR_2022_sifar.pdf>
<papertitle>Can an Image Classifier Suffice For Action Recognition?</papertitle>
</a>
<br>
Quanfu Fan, Chun-Fu (Richard) Chen, <strong>Rameswar Panda</strong><br>
<em> International Conference on Learning Representations (<strong>ICLR</strong>)</em>, 2022 <br>
[<a href="https://github.com/IBM/sifar-pytorch">Code</a>]<br>
<p> We introduce the idea of rearranging input video frames into super images to re-purpose an image classifer for video action recognition.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/NeurIPS_2021_iared.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2021_iared.pdf>
<papertitle>IA-RED<sup>2</sup>: Interpretability-Aware Redundancy Reduction for Vision Transformer</papertitle>
</a>
<br>
Bowen Pan, <strong>Rameswar Panda</strong>, Yifan Jiang, Zhangyang Wang, Rogerio Feris, Aude Oliva<br>
<em>Neural Information Processing Systems (<strong>NeurIPS</strong>)</em>, 2021<br>
[<a href="https://people.csail.mit.edu/bpan/ia-red/">Project Page</a>] [<a href="https://people.csail.mit.edu/bpan/ia-red/">Code</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2021_iared_supp.pdf">Supplementary Material</a>]<br>
<p>We introduce an input-dependent and interpretable dynamic inference framework for vision transformer, which adaptively decides the patch tokens to compute per input instance.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/NeurIPS_2021_comix.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2021_comix.pdf>
<papertitle>Contrast and Mix: Temporal Contrastive Video Domain Adaptation with Background Mixing</papertitle>
</a>
<br>
Aadarsh Sahoo, Rutav Shah, <strong>Rameswar Panda</strong>, Kate Saenko, Abir Das<br>
<em>Neural Information Processing Systems (<strong>NeurIPS</strong>)</em>, 2021<br>
[<a href="https://cvir.github.io/projects/comix">Project Page</a>] [<a href="https://github.com/CVIR/CoMix">Code</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2021_comix_supp.pdf">Supplementary Material</a>]<br>
<p>We propose a new contrastive learning framework that aims to learn discriminative invariant feature representations for unsupervised video domain adaptation.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/NeurIPS_2021_dist.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2021_dist.pdf>
<papertitle>Dynamic Distillation Network for Cross-Domain Few-Shot Recognition with Unlabeled Data</papertitle>
</a>
<br>
Ashraful Islam, Chun-Fu (Richard) Chen, <strong>Rameswar Panda</strong>, Leonid Karlinsky, Rogerio Feris, Richard Radke<br>
<em>Neural Information Processing Systems (<strong>NeurIPS</strong>)</em>, 2021<br>
<em>CVPR Workshop on Learning from Limited or Imperfect Data (<strong>CVPR-W</strong>)</em>, 2021 [<a href="https://rpand002.github.io/data/CVPR_2021_W.pdf">PDF</a>]<br>
[<a href="https://github.com/asrafulashiq/dynamic-cdfsl">Code</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2021_dist_supp.pdf">Supplementary Material</a>]<br>
<p>We address the problem of cross-domain few-shot learning by proposing a simple distillation-based method to facilitate unlabeled images from the novel dataset.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_adamml.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_adamml.pdf>
<papertitle>AdaMML: Adaptive Multi-Modal Learning for Efficient Video Recognition</papertitle>
</a>
<br>
<strong>Rameswar Panda*</strong>, Chun-Fu (Richard) Chen*, Quanfu Fan, Ximeng Sun, Kate Saenko, Aude Oliva, Rogerio Feris<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 <br>
<em>Invited Paper Talk at CVPR Workshop on Sight and Sound (<strong>CVPR-W</strong>)</em>, 2021 [<a href="https://youtu.be/1ReXxKg6vAo?t=2607">Video</a>]<br>
[<a href="https://rpand002.github.io/adamml.html">Project Page</a>] [<a href="https://github.com/IBM/AdaMML">Code</a>] [<a href="https://rpand002.github.io/data/ICCV_2021_adamml_supp.pdf">Supplementary Material</a>]<br>
<p>We propose an adaptive multi-modal learning framework that selects on-the-fly the optimal modalities for each segment conditioned on the input for efficient video recognition.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_videoiq.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_videoiq.pdf>
<papertitle>Dynamic Network Quantization for Efficient Video Inference</papertitle>
</a>
<br>
Ximeng Sun, <strong>Rameswar Panda</strong>, Chun-Fu (Richard) Chen, Aude Oliva, Rogerio Feris, Kate Saenko<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 <br>
[<a href="https://cs-people.bu.edu/sunxm/VideoIQ/project.html">Project Page</a>] [<a href="https://github.com/sunxm2357/VideoIQ">Code</a>] [<a href="https://rpand002.github.io/data/ICCV_2021_videoiq_supp.pdf">Supplementary Material</a>]<br>
<p> We introduce video instance-aware quantization that decides what precision should be used on a per frame basis for efficient video inference.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_crossvit.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_crossvit.pdf>
<papertitle>CrossViT: Cross Attention Multi-Scale Vision Transformer for Image Classification</papertitle>
</a>
<br>
Chun-Fu (Richard) Chen, Quanfu Fan, <strong>Rameswar Panda</strong><br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 (Oral)<br>
[<a href="https://github.com/IBM/CrossViT">Code</a>] [<a href="https://rpand002.github.io/data/ICCV_2021_crossvit_supp.pdf">Supplementary Material</a>]<br>
<p> We develop a dual-branch vision transformer by combining image patches of different sizes to extract multi-scale feature representations for image classification.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_transferability.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_transferability.pdf>
<papertitle>A Broad Study on the Transferability of Visual Representations with Contrastive Learning</papertitle>
</a>
<br>
Ashraful Islam, Chun-Fu (Richard) Chen, <strong>Rameswar Panda</strong>, Leonid Karlinsky, Richard Radke, Rogerio Feris<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 <br>
[<a href="https://github.com/asrafulashiq/transfer_broad">Code</a>] [<a href="https://rpand002.github.io/data/ICCV_2021_transferability_supp.pdf">Supplementary Material</a>]<br>
<p> We conduct extensive analysis on the transferability of contrastive learning on the downstream image classification, few-shot recognition, and object detection tasks.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_mcn.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_mcn.pdf>
<papertitle>Multimodal Clustering Networks for Self-supervised Learning from Unlabeled Videos</papertitle>
</a>
<br>
Brian Chen, Andrew Rouditchenko, Kevin Duarte, Hilde Kuehne, Samuel Thomas, Angie Boggust, <strong>Rameswar Panda</strong>, Brian Kingsbury, Rogerio Feris, David Harwath, James Glass, Michael Picheny, Shih-Fu Chang<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 <br>
<em>CVPR Workshop on Sight and Sound (<strong>CVPR-W</strong>)</em>, 2021 [<a href="data/CVPR_2021_mcn_W.pdf">PDF</a>]<br>
[<a href="https://github.com/brian7685/Multimodal-Clustering-Network">Code</a>] <br>
<p> We extend the concept of instance-level contrastive learning with a multimodal clustering step in the training pipeline to capture semantic similarities across modalities.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICCV_2021_gbs.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2021_gbs.pdf>
<papertitle>Detector-Free Weakly Supervised Grounding by Separation</papertitle>
</a>
<br>
Assaf Arbelle, Sivan Doveh, Amit Alfassy, Joseph Shtok, Guy Lev, Eli Schwartz, Hilde Kuehne, Hila Barak Levi, Prasanna Sattigeri, <strong>Rameswar Panda</strong>, Chun-Fu (Richard) Chen, Alex Bronstein, Kate Saenko, Shimon Ullman, Raja Giryes, Rogerio Feris, Leonid Karlinsky<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2021 (Oral)<br>
<p> We propose a detector-free approach for weakly supervised grounding by learning to separate randomly blended images conditioned on the corresponding texts.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:15%;vertical-align:middle">
<img src="images/ICML_2021.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICML_2021.pdf>
<papertitle>Fair Selective Classification Via Sufficiency</papertitle>
</a>
<br>
Joshua Lee, Yuheng Bu, Deepta Rajan, Prasanna Sattigeri, <strong>Rameswar Panda</strong>, Subhro Das, Gregory Wornell<br>
<em>International Conference on Machine Learning (<strong>ICML</strong>)</em>, 2021 (Oral)<br>
<p> We prove that sufficiency can be used to train fairer selective classifiers which ensure that precision always increases as coverage is decreased for all groups.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2021_Analysis.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2021_Analysis.pdf>
<papertitle>Deep Analysis of CNN-based Spatio-temporal Representations for Action Recognition</papertitle>
</a>
<br>
Chun-Fu (Richard) Chen*, <strong>Rameswar Panda*</strong>, Kandan Ramakrishnan, Rogerio Feris, John Cohn, Aude Oliva, Quanfu Fan*<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2021 <br>
[<a href="https://github.com/IBM/action-recognition-pytorch">Code</a>] [<a href="https://rpand002.github.io/data/CVPR_2021_Analysis_Supp.pdf">Supplementary Material</a>]<br>
<p> We develop a unified framework for action recognition models and systematically compare them to better understand the differences and spatio-temporal behavior on large-scale datasets.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2021_TCL.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2021_TCL.pdf>
<papertitle>Semi-Supervised Action Recognition with Temporal Contrastive Learning</papertitle>
</a>
<br>
Ankit Singh, Omprakash Chakraborty, Ashutosh Varshney, <strong>Rameswar Panda</strong>, Rogerio Feris, Kate Saenko, Abir Das<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2021 <br>
[<a href="https://cvir.github.io/TCL/">Project Page</a>] [<a href="https://github.com/CVIR/TCL/">Code</a>] [<a href="https://rpand002.github.io/data/CVPR_2021_TCL_Supp.pdf">Supplementary Material</a>]<br>
<p> We propose a temporal contrastive learning framework for semi-supervised action recognition by using contrastive losses between different videos and groups of videos with similar actions.</p>
</td>
</tr> </tbody></table> <br>
<!-- <table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>-->
<!-- <tr onmouseout="nlt_stop()" onmouseover="nlt_start()">-->
<!-- <td style="padding:20px;width:25%;vertical-align:middle">-->
<!-- <img src="images/CVPR_2021_W.png" width="160" height="100">-->
<!-- </td>-->
<!-- <td width="75%" valign="middle">-->
<!-- <a href=data/CVPR_2021_W.pdf>-->
<!-- <papertitle>A Simple Framework for Cross-Domain Few-Shot Recognition with Unlabeled Data</papertitle>-->
<!-- </a>-->
<!-- <br>-->
<!-- Ashraful Islam, Richard Chen, <strong>Rameswar Panda</strong>, Leonid Karlinsky, Rogerio Feris, Richard Radke<br>-->
<!-- <em>CVPR Workshop on Learning from Limited or Imperfect Data (<strong>CVPR-W</strong>)</em>, 2021 <br>-->
<!-- <p> We introduce a simple pre-training framework to utilize unlabeled data from the target domain for cross-domain few-shot learning.</p>-->
<!-- </td>-->
<!-- </tr> </tbody></table> <br>-->
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICLR_2021_AdaFuse.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICLR_2021_AdaFuse.pdf>
<papertitle>AdaFuse: Adaptive Temporal Fusion Network for Efficient Action Recognition</papertitle>
</a>
<br>
Yue Meng, <strong>Rameswar Panda</strong>, Chung-Ching Lin, Prasanna Sattigeri, Leonid Karlinsky, Kate Saenko, Aude Oliva, Rogerio Feris<br>
<em> International Conference on Learning Representations (<strong>ICLR</strong>)</em>, 2021 <br>
[<a href="https://mengyuest.github.io/AdaFuse/">Project Page</a>] [<a href="https://github.com/mengyuest/AdaFuse">Code</a>]<br>
<p> We introduce an adaptive temporal fusion network that dynamically fuses channels from current and past feature maps for strong temporal modelling in action recognition.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICLR_2021_VARED.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ICLR_2021_VARED.pdf>
<papertitle>VA-RED<sup>2</sup>: Video Adaptive Redundancy Reduction</papertitle>
</a>
<br>
Bowen Pan, <strong>Rameswar Panda</strong>, Camilo Fosco, Chung-Ching Lin, Alex Andonian, Yue Meng, Kate Saenko, Aude Oliva, Rogerio Feris<br>
<em> International Conference on Learning Representations (<strong>ICLR</strong>)</em>, 2021 <br>
[<a href="http://people.csail.mit.edu/bpan/va-red/">Project Page</a>] [<a href="http://people.csail.mit.edu/bpan/va-red/">Code</a>]<br>
<p> We propose an input-dependent adaptive framework for efficient video recognition that automatically decides what feature maps to compute per input instance.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/AAAI_2021.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/AAAI_2021_NAS.pdf>
<papertitle>NASTransfer: Analyzing Architecture Transferability in Large Scale Neural Architecture Search</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Michele Merler, Mayoore Jaiswal, Hui Wu, Kandan Ramakrishnan, Ulrich Finkler,
Richard Chen, Minsik Cho, Rogerio Feris, David Kung, Bishwaranjan Bhattacharjee <br>
<em> AAAI Conference on Artificial Intelligence (<strong>AAAI</strong>)</em>, 2021 <br>
<p> We analyze the architecture transferability of different NAS methods by performing a series of experiments on several large scale image benchmarks.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/AAAI_2021_W.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/AAAI_2021_W.pdf>
<papertitle>Large Scale Neural Architecture Search with Polyharmonic Splines</papertitle>
</a>
<br>
Ulrich Finkler, Michele Merler, <strong>Rameswar Panda</strong>, Mayoore Jaiswal, Hui Wu, Kandan Ramakrishnan,
Chun-Fu Chen, Minsik Cho, David Kung, Rogerio Feris, Bishwaranjan Bhattacharjee <br>
<em> AAAI Workshop on Meta-Learning for Computer Vision (<strong>AAAI-W</strong>)</em>, 2021 <br>
<p> We propose a NAS method based on polyharmonic splines that can perform architecture search directly on large scale image datasets like ImageNet22K.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/NeurIPS_2020.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPS_2020.pdf>
<papertitle>AdaShare: Learning What To Share For Efficient Deep Multi-Task Learning</papertitle>
</a>
<br>
Ximeng Sun, <strong>Rameswar Panda</strong>, Rogerio Feris, Kate Saenko <br>
<em> Neural Information Processing Systems (<strong>NeurIPS</strong>)</em>, 2020 <br>
[<a href="https://cs-people.bu.edu/sunxm/AdaShare/project.html">Project Page</a>] [<a href="https://github.com/sunxm2357/AdaShare">Code</a>] [<a href="https://rpand002.github.io/data/NeurIPS_2020_Supp.pdf">Supplementary Material</a>]<br>
<p>We propose a novel approach for adaptively determining the feature sharing pattern across multiple tasks (what layers to share across which tasks) in deep multi-task learning.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/TCSVT_2020.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/TCSVT_2020.pdf>
<papertitle>Exploiting Global Camera Network Constraints for Unsupervised Video Person Re-identification</papertitle>
</a>
<br>
Xueping Wang, <strong>Rameswar Panda</strong>, Min Liu, Yaonan Wang, Amit K. Roy-Chowdhury<br>
<em>IEEE Transactions on Circuits and Systems for Video Technology (<strong>TCSVT</strong>)</em>, 2020
<p> We propose a consistent cross-view matching framework, in which global camera network constraints are exploited to address the problem of unsupervised video-based re-identification.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/MM_2020.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/MM_2020.pdf>
<papertitle>Adversarial Knowledge Transfer from Unlabeled Data</papertitle>
</a>
<br>
Akash Gupta*, <strong>Rameswar Panda*</strong>, Sujoy Paul, Jianming Zhang, Amit K. Roy-Chowdhury<br>
<em>ACM Multimedia (<strong>MM</strong>)</em>, 2020 <br>
[<a href="https://agupt013.github.io/akt.html">Project Page</a>] [<a href="https://github.com/agupt013/akt">Code</a>] <br>
<p>We present a novel adversarial framework for transferring knowledge from internet-scale unlabeled data to improve the performance of a classifier on a given visual recognition task.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ECCV_2020.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/ECCV_2020.pdf>
<papertitle>AR-Net: Adaptive Frame Resolution for Efficient Action Recognition</papertitle>
</a>
<br>
Yue Meng, Chung-Ching Lin, <strong>Rameswar Panda</strong>, Prasanna Sattigeri, Leonid Karlinsky, Aude Oliva, Kate Saenko, Rogerio Feris <br>
<em>European Conference on Computer Vision (<strong>ECCV</strong>)</em>, 2020 <br>
[<a href="https://mengyuest.github.io/AR-Net/">Project Page</a>] [<a href="https://github.com/mengyuest/AR-Net">Code</a>] <br>
<p>We propose an adaptive approach to select optimal resolution for each frame conditioned on the input for efficient action recognition in long untrimmed video.</p>
</td>
</tr>
</tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ECCV_2020_W.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/ECCV_2020_W.pdf>
<papertitle>Mitigating Dataset Imbalance via Joint Generation and Classification</papertitle>
</a>
<br>
Aadarsh Sahoo, Ankit Singh, <strong>Rameswar Panda</strong>, Rogerio Feris, Abir Das<br>
<em>ECCV Workshop on Imbalance Problems in Computer Vision (<strong>ECCV-W</strong>)</em>, 2020
<p>We introduce a joint dataset repairment strategy by combining classifier with a GAN that makes up for the deficit of training examples from the minority class by producing additional examples.</p>
</td>
</tr>
</tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/MICCAI_2020.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/MICCAI_2020.pdf>
<papertitle>Fairness of Classifiers Across Skin Tones in Dermatology</papertitle>
</a>
<br>
Newton M. Kinyanjui, Timothy Odonga, Celia Cintas, Noel C. F. Codella, <strong>Rameswar Panda</strong>, Prasanna Sattigeri, Kush R. Varshney<br>
<em>Medical Image Computing and Computer Assisted Interventions (<strong>MICCAI</strong>)</em>, 2020
<p>We present an approach to estimate the consistency in performance of classifiers across populations with varying skin tones in skin disease benchmarks.</p>
</td>
</tr>
</tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2020_Video.jpg" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2020_Video.pdf>
<papertitle>Non-Adversarial Video Synthesis with Learned Priors</papertitle>
</a>
<br>
Abhishek Aich, Akash Gupta, <strong>Rameswar Panda</strong>, Rakib Hyder, Salman Asif, Amit K. Roy-Chowdhury<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2020 <br>
[<a href="https://abhishekaich27.github.io/navsynth.html">Project Page</a>] [<a href="https://github.com/abhishekaich27/Navsynth">Code</a>] <br>
<p>We introduce a novel non-adversarial framework for generating a wide range of diverse videos from latent noise vectors without any any conditional input reference frame.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2020_ReID.jpg" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2020_ReID.pdf>
<papertitle>Camera On-boarding for Person Re-identification using Hypothesis Transfer Learning</papertitle>
</a>
<br>
Sk Miraj Ahmed, Aske R. Lejbolle, <strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2020
<p>We propose an approach to swiftly on-board new camera(s) in an existing re-id network using only source models and limited labeled data, but without having access to source camera data.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2020_W.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2020_W.pdf>
<papertitle>Relationship Matters: Relation Guided Knowledge Transfer for Incremental Learning of Object Detectors</papertitle>
</a>
<br>
Kandan Ramakrishnan, <strong>Rameswar Panda</strong>, Quanfu Fan, John Henning, Aude Oliva, Rogerio Feris<br>
<em>CVPR Workshop on Continual Learning in Computer Vision (<strong>CVPR-W</strong>)</em>, 2020
<p>We introduce a novel approach that focuses on object relations to effectively transfer knowledge for minimizing the effect of catastrophic forgetting in incremental learning of object detectors.</p>
</td>
</tr> </tbody></table> <br>
<!-- <div style="width: 100%; height: 20px; border-bottom: 1px solid black; text-align: center">-->
<!-- <span style="font-size: 20px; background-color: #F3F5F6; padding: 0 10px;">-->
<!-- 2019 <!–Padding is optional–>-->
<!-- </span>-->
<!-- </div> <br>-->
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/PR_2019.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/PR_2019.pdf>
<papertitle>Adaptation of Person Re-identification Models for On-boarding New Camera(s)</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Amran Bhuiyan, Vittorio Murino, Amit K. Roy-Chowdhury<br>
<em>Pattern Recognition (<strong>PR</strong>)</em>, 2019
<p>This paper extends our CVPR 2017 paper providing a new source-target selective adaptation strategy and rigorous experiments on more person re-id datasets.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/TCSVT_2019.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/TCSVT_2019.pdf>
<papertitle>Construction of Diverse Image Datasets from Web Collections with Limited Labeling</papertitle>
</a>
<br>
Niluthpol C. Mithun, <strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>IEEE Transactions on Circuits and Systems for Video Technology (<strong>TCSVT</strong>)</em>, 2019
<p>This paper extends our MM 2016 paper where we employ a joint visual-semantic space to simultaneously utilize both images and text from the web for dataset construction.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/NeurIPSW_2019.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/NeurIPSW_2019.pdf>
<papertitle>Estimating Skin Tone and Effects on Classification Performance in Dermatology Datasets</papertitle>
</a>
<br>
Newton M. Kinyanjui, Timothy Odonga, Celia Cintas, Noel C. F. Codella, <strong>Rameswar Panda</strong>, Prasanna Sattigeri, Kush R. Varshney<br>
<em>NeurIPS Fair Machine Learning for Health Workshop (<strong>NeurIPS-W</strong>)</em>, 2019
<p>We present an approach to estimate skin tone in benchmark skin disease datasets, and investigate whether model performance is dependent on this measure</p>
</td>
</tr>
</tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ECCV_2018.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/ECCV_2018.pdf>
<papertitle>Contemplating Visual Emotions: Understanding and Overcoming Dataset Bias</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Jianming Zhang, Haoxiang Li, Joon-Young Lee, Xin Lu, Amit K. Roy-Chowdhury<br>
<em>European Conference on Computer Vision (<strong>ECCV</strong>)</em>, 2018 <br>
[<a href="https://rpand002.github.io/emotion.html">Project Page</a>] [<a href="https://rpand002.github.io/data/ECCV_2018_Supp.pdf">Supplementary Material</a>]<br>
<p>We investigate different dataset biases and propose a curriculum guided webly supervised approach for learning a generalizable emotion recognition model.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/MM_2018.png" width="160" height="100">
</td>
<td width="75%" valign="middle">
<a href=data/MM_2018.pdf>
<papertitle>Webly Supervised Joint Embedding for Cross-Modal Image-Text Retrieval</papertitle>
</a>
<br>
Niluthpol C. Mithun, <strong>Rameswar Panda</strong>, Evangelos E. Papalexakis, Amit K. Roy-Chowdhury<br>
<em>ACM Multimedia (<strong>MM</strong>)</em>, 2018
<p>This work exploits large scale web data for learning an effective multi-modal embedding without requiring large amount of human-crafted training data.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2018.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2018.pdf>
<papertitle>FFNet: Video Fast-Forwarding via Reinforcement Learning</papertitle>
</a>
<br>
Shuyue Lan, <strong>Rameswar Panda</strong>, Qi Zhu, Amit K. Roy-Chowdhury<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2018
<p>We introduce an online framework for fast-forwarding a video while presenting its important and interesting content on the fly without processing or even obtaining the entire video.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICCV_2017.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/ICCV_2017.pdf>
<papertitle>Weakly Supervised Summarization of Web Videos</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Abir Das, Ziyan Wu, Jan Ernst, Amit K. Roy-Chowdhury<br>
<em>International Conference on Computer Vision (<strong>ICCV</strong>)</em>, 2017
<p>We introduce a weakly supervised approach that requires only video-level annotations for summarizing long unconstrained web videos.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2017_ReID.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2017_RD.pdf>
<papertitle>Unsupervised Adaptive Re-identification in Open World Dynamic Camera Networks</papertitle>
</a>
<br>
<strong>Rameswar Panda*</strong>, Amran Bhuiyan*, Vittorio Murino, Amit K. Roy-Chowdhury<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2017
<p>We propose an unsupervised adaptation scheme for re-identification models where a new camera may be temporarily inserted into an existing system to get additional information.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVPR_2017_Summ.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/CVPR_2017_Summ.pdf>
<papertitle>Collaborative Summarization of Topic-Related Videos</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>IEEE Conference on Computer Vision and Pattern Recognition (<strong>CVPR</strong>)</em>, 2017
<p>This paper presents a collaborative video summarization approach that exploits visual context from a set of topic-related videos to extract an informative summary of a given video.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/TMM_2017.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/TMM_2017.pdf>
<papertitle>Multi-View Surveillance Video Summarization via Joint Embedding and Sparse Optimization</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>IEEE Transactions on Multimedia (<strong>TMM</strong>)</em>, 2017
<p>This paper extends our ICPR 2016 paper providing new theoretical insights with a joint optimization and experimenting on spatio-temporal features and datasets.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/TIP_2017.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/TIP_2017.pdf>
<papertitle>Diversity-aware Multi-Video Summarization</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Niluthpol C. Mithun, Amit K. Roy-Chowdhury<br>
<em>IEEE Transactions on Image Processing (<strong>TIP</strong>)</em>, 2017
<p>This paper introduces a new generalized sparse optimization framework for summarizing multiple videos generated from a video search or from a multi-view camera network.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/TCYB_2017.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/TCYB_2017.pdf>
<papertitle>Nystrom approximated temporally constrained multi-similarity spectral clustering approach for movie scene detection</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Sanjay K. Kuanar, Ananda S. Chowdhury<br>
<em>IEEE Transactions on Cybernetics (<strong>TCYB</strong>)</em>, 2017
<p>We present a fast solution for movie scene detection using Nystrom approximated multi-similarity spectral clustering with a temporal integrity constraint.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICASSP_2017.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/ICASSP_2017.pdf>
<papertitle>Sparse Modeling for Topic-oriented Video Summarization</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>IEEE International Conference on Acoustics, Speech and Signal Processing (<strong>ICASSP</strong>)</em>, 2017
<p>This paper presents a diversity-aware sparse optimization framework for summarizing topi-related videos generated from a video search.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/CVIU_2016.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/CVIU_2016.pdf>
<papertitle>Continuous Adaptation of Multi-Camera Person Identification Models through Sparse Non-redundant Representative Selection</papertitle>
</a>
<br>
Abir Das, <strong>Rameswar Panda</strong>, Amit K. Roy-Chowdhury<br>
<em>Computer Vision and Image Understanding (<strong>CVIU</strong>)</em>, 2016
<p>We addressed the problem of online learning of identification systems where unlabeled data comes in small minibatches, with human in the loop.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICPR_2016.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/ICPR_2016.pdf>
<papertitle>Video Summarization in a Multi-View Camera Network</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Abir Das, Amit K. Roy-Chowdhury<br>
<em>IEEE International Conference on Pattern Recognition (<strong>ICPR</strong>)</em>, 2016
<p>This paper presents a framework for summarizing multi-view videos by exploiting both intra- and inter-view content correlations in a joint embedding space.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/ICIP_2016.png" width="160">
</td>
<td width="75%" valign="middle">
<a href=data/ICIP_2016.pdf>
<papertitle>Embedded Sparse Coding for Summarizing Multi-View Videos</papertitle>
</a>
<br>
<strong>Rameswar Panda</strong>, Abir Das, Amit K. Roy-Chowdhury<br>
<em>IEEE International Conference on Image Processing (<strong>ICIP</strong>)</em>, 2016
<p>This paper presents a stochastic multi-view frame embedding based on KL divergence to preserve correlations in multi-view learning.</p>
</td>
</tr> </tbody></table> <br>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr onmouseout="nlt_stop()" onmouseover="nlt_start()">
<td style="padding:20px;width:25%;vertical-align:middle">
<img src="images/MM_2016.png" width="160" height="90">
</td>
<td width="75%" valign="middle">
<a href=data/MM_2016.pdf>
<papertitle>Generating Diverse Image Datasets with Limited Labeling</papertitle>
</a>
<br>