forked from academicpages/academicpages.github.io
-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathindex.html
625 lines (534 loc) · 36 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
<!DOCTYPE HTML "-//W3C//DTD HTML 4.01 Transitional//EN">
<html lang="en">
<head>
<!-- Google tag (gtag.js) -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-GLG5GQ8E08"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-GLG5GQ8E08');
</script>
<meta name="generator" content="HTML Tidy for Linux/x86 (vers 11 February 2007), see www.w3.org">
<style type="text/css">
/* Design Credits: Jon Barron, Deepak Pathak*/
a {
color: #1772d0;
text-decoration: none;
}
a:focus,
a:hover {
color: #f09228;
text-decoration: none;
}
body,
td,
th {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
font-size: 16px;
font-weight: 400
}
heading {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
font-size: 19px;
font-weight: 1000
}
strong {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
font-size: 16px;
font-weight: 800
}
strongred {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
color: 'red';
font-size: 16px
}
sectionheading {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
font-size: 22px;
font-weight: 600
}
pageheading {
font-family: 'Titillium Web', Verdana, Helvetica, sans-serif;
font-size: 38px;
font-weight: 400
}
</style>
<!-- <link rel="icon" type="image/png" href="images/W.png"> -->
<script type="text/javascript" src="js/hidebib.js"></script>
<title>Shubhang Bhatnagar</title>
<meta name="author" content="Shubhang Bhatnagar">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
<link rel="icon" href="data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 100 100%22><text y=%22.9em%22 font-size=%2290%22>🌐</text></svg>">
</head>
<body>
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<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;">
<tr style="padding:0px">
<td style="padding:2.5%;width:60%;vertical-align:middle">
<p class="name" style="text-align: center;">
Shubhang Bhatnagar
</p>
<p>Hi, I am Shubhang Bhatnagar, a fourth year PhD student in the <a href="https://vision.ai.illinois.edu/">Computer Vision and Robotics Laboratory</a> at the <a href="https://illinois.edu/">University of Illinois Urbana-Champaign</a>.
I am grateful to be advised by <a href="https://vision.ai.illinois.edu/narendra-ahuja/"> Prof. Narendra Ahuja </a>.
</p>
<p>
Broadly, I am interested in computer vision, machine learning and their applications.
<!-- Currently, my research focuses on semantic representation learning for robust, zero-shot recognition and retrieval.-->
</p>
<p>Previously, I completed my Dual degree (B.Tech + M.Tech) in electrical engineering from <a href="https://www.iitb.ac.in/">Indian Institute of Technology, Bombay</a>, where I was awarded the Institute Silver medal for graduating at the top of my batch.
In my Master's thesis, I worked on developing label efficient deep learning techniques advised by <a href="https://www.ee.iitb.ac.in/~asethi/">Prof. Amit Sethi</a>.
</p>
<p style="text-align:center">
<a target="_blank" href="mailto:[email protected]">E-mail</a> /
<a target="_blank" href="https://shubhangb97.github.io/files/Resume_Shubhang.pdf">Resume</a> /
<a href="https://github.com/shubhangb97">GitHub</a> /
<a href="https://scholar.google.com/citations?user=x_6C2vEAAAAJ&hl=en">Google Scholar</a> /
<a href="https://www.linkedin.com/in/shubhang-bhatnagar-b78b24186"> LinkedIn </a> /
<a href="https://twitter.com/s_bhatnagar_tw">Twitter</a>
</p>
</td>
<td style="padding:2.5%;width:40%;max-width:40%">
<img style="width:100%;max-width:100%" alt="profile photo" src="images/Web_img1.jpg">
</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">
<h2>Research</h2>
</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:40%;vertical-align:top;min-width:120px">
<a href="https://shubhangb97.github.io/potential_field_DML"><img src="images/charge_animation_small.gif" alt="project image" style="width:auto; height:auto; max-width:100%;" /> </a>
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Potential Field Based Deep Metric Learning</h3>
<br>
<strong>Shubhang Bhatnagar</strong>, <a href="https://en.wikipedia.org/wiki/Narendra_Ahuja"> Narendra Ahuja </a>
<br>
<em>Under Review </em>
<br>
<a href="javascript:toggleblock('potential_field_abs')">abstract</a> /
<a href="https://shubhangb97.github.io/potential_field_DML">project page</a> /
<a href="https://arxiv.org/abs/2405.18560">arxiv preprint</a>
<br>
<p align="justify"> <i id="potential_field_abs">Deep metric learning (DML) involves training a network to learn a semantically meaningful representation space.
Many current approaches mine n-tuples of examples and model interactions within each tuplets. We present a novel, compositional DML model,
inspired by electrostatic fields in physics that, instead of in tuples, represents the influence of each example (embedding) by a continuous
potential field, and superposes the fields to obtain their combined global potential field. We use attractive/repulsive potential fields to represent interactions
among embeddings from images of the same/different classes. Contrary to typical learning methods, where mutual influence of samples is proportional to their distance,
we enforce reduction in such influence with distance, leading to a decaying field. We show that such decay helps improve performance on real world datasets with large
intra-class variations and label noise. Like other proxy-based methods, we also use proxies to succinctly represent sub-populations of examples. We evaluate our method
on three standard DML benchmarks- Cars-196, CUB-200-2011, and SOP datasets where it outperforms state-of-the-art baselines.</i></p>
</td>
</tr>
<tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/positive_coop_snap.png" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>PositiveCoOp: Rethinking Prompting Strategies for Multi-Label Recognition with Partial Annotations</h3>
<br>
<a href="https://samyakr99.github.io">Samyak Rawlekar</a>,<strong>Shubhang Bhatnagar</strong>, <a href="https://en.wikipedia.org/wiki/Narendra_Ahuja"> Narendra Ahuja </a>
<br>
<em> <a href="https://wacv2025.thecvf.com/">WACV </a></em> 2025,
<br>
<a href="javascript:toggleblock('MLR2_abs')">abstract</a> /
<a href="https://samyakr99.github.io/PositiveCoOp/">project page</a> /
<a href="https://arxiv.org/pdf/2409.08381">paper</a>
<!-- <a href="coming soon">video</a> /
-->
<br>
<p align="justify"> <i id="MLR2_abs"> Vision-language models (VLMs) like CLIP have been adapted for Multi-Label Recognition (MLR) with partial annotations by leveraging prompt-learning, where positive and negative prompts
are learned for each class to associate their embeddings with class presence or absence in the shared vision-text feature space. While this approach improves MLR performance by
relying on VLM priors, we hypothesize that learning negative prompts may be suboptimal, as the datasets used to train VLMs lack image-caption pairs explicitly focusing on class
absence. To analyze the impact of positive and negative prompt learning on MLR, we introduce PositiveCoOp and NegativeCoOp, where only one prompt is learned with VLM guidance
while the other is replaced by an embedding vector learned directly in the shared feature space without relying on the text encoder. Through empirical analysis, we observe that
negative prompts degrade MLR performance, and learning only positive prompts, combined with learned negative embeddings (PositiveCoOp), outperforms dual prompt learning approaches.
Moreover, we quantify the performance benefits that prompt-learning offers over a simple vision-features-only baseline, observing that the baseline displays strong performance
comparable to dual prompt learning approach (DualCoOp), when the proportion of missing labels is low, while requiring half the training compute and 16 times fewer parameters. </i></p>
</td>
</tr>
<tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/MLR1_snap.PNG" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Improving Multi-label Recognition using Class Co-Occurrence Probabilities</h3>
<br>
<strong>Shubhang Bhatnagar*</strong>, <a href="https://samyakr99.github.io">Samyak Rawlekar* </a>, Vishnuvardhan Pogunulu Srinivasulu, <a href="https://en.wikipedia.org/wiki/Narendra_Ahuja"> Narendra Ahuja </a>
<br>
<em> <a href="https://sites.google.com/view/cvpr-metafood-2024">CVPRW </a></em> 2024,<em> <a href="https://icpr2024.org/">ICPR </a> 2024 (Oral Top-5%)</em>
<br>
<a href="javascript:toggleblock('MLR1_abs')">abstract</a> /
<a href="https://shubhangb97.github.io/MLR_gcn">project page</a> /
<a href="https://arxiv.org/abs/2404.16193">paper</a>
<!-- <a href="coming soon">video</a> /
-->
<br>
<p align="justify"> <i id="MLR1_abs">Multi-label Recognition (MLR) involves the identification of multiple objects within an image.
To address the additional complexity of this problem, recent works have leveraged information from vision-language models (VLMs) trained on large text-images datasets for
the task. These methods learn an independent classifier for each object (class), overlooking correlations in their occurrences. Such co-occurrences can be captured from the
training data as conditional probabilities between a pair of classes. We propose a framework to extend the independent classifiers by incorporating the
co-occurrence information for object pairs to improve the performance of independent classifiers. We use a Graph Convolutional Network (GCN) to enforce the
conditional probabilities between classes, by refining the initial estimates derived from image and text sources obtained using VLMs. We validate our method on four MLR
datasets, where our approach outperforms all state-of-the-art methods.</i></p>
</td>
</tr>
<tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/Piecewise_manifold_UDML.png" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Piecewise-Linear Manifolds for Deep Metric Learning</h3>
<br>
<strong>Shubhang Bhatnagar</strong>, <a href="https://vision.ai.illinois.edu/narendra-ahuja/"> Narendra Ahuja </a>
<br>
<em> <a href="https://proceedings.mlr.press/v234/">In Proceedings of Machine Learning Research (PMLR) Vol. 234, </a></em>
<em> <a href="https://cpal.cc/">Conference on Parsimony and Learning (CPAL) </a></em>, 2024 <b>(Oral) </b>
<br>
<a href="javascript:toggleblock('PL_UDML_abs')">abstract</a> /
<a href="https://shubhangb97.github.io/PL_manifold_UDML">project page</a> /
<a href="https://proceedings.mlr.press/v234/bhatnagar24a.html">paper</a>
<!-- <a href="coming soon">video</a> /
-->
<br>
<p align="justify"> <i id="PL_UDML_abs"> Unsupervised deep metric learning (UDML) focuses on learning a semantic representation space
using only unlabeled data. This challenging problem requires accurately estimating the similarity between data points, which is used
to supervise a deep network. For this purpose, we propose to model the high-dimensional data manifold using a piecewise-linear
approximation, with each low-dimensional linear piece approximating the data manifold in a small neighborhood of a point.
These neighborhoods are used to estimate similarity between data points. We empirically show that this similarity estimate correlates
better with the ground truth than the similarity estimates of current state-of-the-art techniques. We also show that proxies, commonly
used in supervised metric learning, can be used to model the piecewise-linear manifold in an unsupervised setting, helping improve
performance. Our method outperforms existing unsupervised metric learning approaches on standard zero-shot image retrieval benchmarks.</i></p>
</td>
</tr>
<!--<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Rethinking Prompting Strategies for Multi-Label Recognition with Partial
Annotations</h3>
<br>
<strong>Shubhang Bhatnagar</strong>, <a href="https://vision.ai.illinois.edu/narendra-ahuja/"> Narendra Ahuja </a>
<br>
<em> <a href="https://proceedings.mlr.press/v234/">In Proceedings of Machine Learning Research (PMLR) Vol. 234, </a></em>
<em> <a href="https://cpal.cc/">Conference on Parsimony and Learning (CPAL) </a></em>, 2024 <b>(Oral) </b>
<br>
<a href="javascript:toggleblock('PL_UDML_abs')">abstract</a> /
<a href="https://shubhangb97.github.io/PL_manifold_UDML">project page</a> /
<a href="https://proceedings.mlr.press/v234/bhatnagar24a.html">paper</a>
<!-- <a href="coming soon">video</a> /
<br>
<p align="justify"> <i id="PL_UDML_abs"> Unsupervised deep metric learning (UDML) focuses on learning a semantic representation space
using only unlabeled data. This challenging problem requires accurately estimating the similarity between data points, which is used
to supervise a deep network. For this purpose, we propose to model the high-dimensional data manifold using a piecewise-linear
approximation, with each low-dimensional linear piece approximating the data manifold in a small neighborhood of a point.
These neighborhoods are used to estimate similarity between data points. We empirically show that this similarity estimate correlates
better with the ground truth than the similarity estimates of current state-of-the-art techniques. We also show that proxies, commonly
used in supervised metric learning, can be used to model the piecewise-linear manifold in an unsupervised setting, helping improve
performance. Our method outperforms existing unsupervised metric learning approaches on standard zero-shot image retrieval benchmarks.</i></p>
</td>
</tr>-->
<tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/ld_gr_iros.png" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Long-Distance Gesture Recognition using Dynamic Neural Networks</h3>
<br>
<strong>Shubhang Bhatnagar</strong>, <a href="https://www.linkedin.com/in/sharathkumargopal"> Sharath Gopal </a>, <a href="https://vision.ai.illinois.edu/narendra-ahuja/"> Narendra Ahuja </a>, <a href="https://sites.google.com/site/liurenshomepage/">Liu Ren</a>
<br>
<em>International Conference on Intelligent Robots and Systems (IROS) </em>, 2023
<br>
<a href="javascript:toggleblock('ldgest_abs')">abstract</a> /
<a href="https://shubhangb97.github.io/LD_GR_dynamic_nn">project page</a> /
<a href="https://ieeexplore.ieee.org/document/10342147">paper</a> /
<a href="https://arxiv.org/abs/2308.04643">arxiv</a>
<!-- <a href="coming soon">video</a> /
-->
<br>
<p align="justify"> <i id="ldgest_abs">Gestures form an important medium of communication between humans and machines. An overwhelming
majority of existing gesture recognition methods are tailored to
a scenario where humans and machines are located very close
to each other. This short-distance assumption does not hold
true for several types of interactions, for example gesture-based
interactions with a floor cleaning robot or with a drone. Methods
made for short-distance recognition are unable to perform
well on long-distance recognition due to gestures occupying
only a small portion of the input data. Their performance is
especially worse in resource constrained settings where they
are not able to effectively focus their limited compute on the
gesturing subject. We propose a novel, accurate and efficient
method for the recognition of gestures from longer distances. It
uses a dynamic neural network to select features from gesturecontaining
spatial regions of the input sensor data for further
processing. This helps the network focus on features important
for gesture recognition while discarding background features
early on, thus making it more compute efficient compared
to other techniques. We demonstrate the performance of our
method on the LD-ConGR long-distance dataset where it
outperforms previous state-of-the-art methods on recognition
accuracy and compute efficiency.</i></p>
</td>
</tr>
<tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/PAL_bmvc.png" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>PAL: Pretext based Active Learning</h3>
<br>
<strong>Shubhang Bhatnagar</strong>, <a href="https://saching007.github.io/">Sachin Goyal</a>, Darshan Tank, <a href="https://www.ee.iitb.ac.in/~asethi/">Amit Sethi</a>
<br><strong>British Machine Vision Conference (BMVC), 2021</strong>
<br>
<a href="javascript:toggleblock('pal_abs')">abstract</a> /
<a href="https://www.bmvc2021-virtualconference.com/conference/papers/paper_1061.html">project page</a> /
<a href="https://www.bmvc2021-virtualconference.com/assets/papers/1061.pdf">paper</a> /
<a href="https://github.com/shubhangb97/PAL_pretext_based_active_learning">code</a>
<br>
<p align="justify"> <i id="pal_abs">The goal of pool-based active learning is to judiciously select a fixed-sized subset of
unlabeled samples from a pool to query an oracle for their labels, in order to maximize
the accuracy of a supervised learner. However, the unsaid requirement that the oracle
should always assign correct labels is unreasonable for most situations. We propose an
active learning technique for deep neural networks that is more robust to mislabeling than
the previously proposed techniques. Previous techniques rely on the task network itself
to estimate the novelty of the unlabeled samples, but learning the task (generalization)
and selecting samples (out-of-distribution detection) can be conflicting goals. We use
a separate network to score the unlabeled samples for selection. The scoring network
relies on self-supervision for modeling the distribution of the labeled samples to reduce
the dependency on potentially noisy labels. To counter the paucity of data, we also deploy
another head on the scoring network for regularization via multi-task learning and use an
unusual self-balancing hybrid scoring function. Furthermore, we divide each query into
sub-queries before labeling to ensure that the query has diverse samples. In addition to
having a higher tolerance to mislabeling of samples by the oracle, the resultant technique
also produces competitive accuracy in the absence of label noise. The technique also
handles the introduction of new classes on-the-fly well by temporarily increasing the
sampling rate of these classes. We make our code publicly available at https://
github.com/shubhangb97/PAL_pretext_based_active_learning</i></p>
</td>
</tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/l1_eusipco.png" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Analyzing Cross Validation in Compressed Sensing with Mixed Gaussian and Impulse Measurement Noise with L1 Errors</h3>
<br>
<strong>Shubhang Bhatnagar</strong><sup>*</sup>, Chinmay Gurjarpadhye<sup>*</sup>, <a href="https://www.cse.iitb.ac.in/~ajitvr/">Ajit Rajwade </a>
<br><strong>European Signal Processing Conference (EUSIPCO),</strong>
<br>
<a href="javascript:toggleblock('l1_cv_abs')">abstract</a> /
<a href="https://ieeexplore.ieee.org/document/9615951">paper</a> /
<a href="https://arxiv.org/abs/2102.10165">extended arxiv</a>
<br>
<p align="justify"> <i id="l1_cv_abs">Compressed sensing (CS) involves sampling
signals at rates less than their Nyquist rates and attempting to reconstruct them after sample acquisition.
Most such algorithms have parameters, for example the regularization parameter in LASSO, which need to be chosen carefully for optimal
performance. These parameters can be chosen based on assumptions on the noise level or signal sparsity, but this knowledge may often be
unavailable. In such cases, cross validation (CV) can be used to choose these parameters in a purely data-driven fashion. Previous work
analyzing the use of CV in CS has been based on the ℓ2 cross-validation error with Gaussian measurement noise. But it is well known that the ℓ2
error is not robust to impulse noise and provides a poor estimate of the recovery error, failing to choose the best parameter. Here we propose
using the ℓ1−CV error which provides substantial performance benefits given impulse measurement noise. Most importantly, we provide a detailed
theoretical analysis and error bounds for the use of ℓ1−CV error in CS reconstruction. We show that with high probability, choosing the parameter
that yields the minimum ℓ1−CV error is equivalent to choosing the minimum recovery error (which is not observable in practice). To our best
knowledge, this is the first paper which theoretically analyzes ℓ1 -based CV in CS.</i></p>
</td>
</tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/subband_coder.PNG" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Insights on coding gain and its properties for principal component filter banks</h3>
<br>
Prasad Chaphekar, Aniket Bhatia,<strong>Shubhang Bhatnagar</strong>, Abhiraj Kanse, Ashish V Vanmali, <a href="https://www.ee.iitb.ac.in/wiki/faculty/vmgadre">Vikram M Gadre </a>
<br><strong> Sādhanā</strong> , Journal of the Indian Academy of Sciences, 2023
<br>
<a href="javascript:toggleblock('pcfb_abs')">abstract</a> /
<a href="https://rdcu.be/dbulY">paper</a>
<br>
<p align="justify"> <i id="pcfb_abs">Principal Component Filter Bank (PCFB) is considered optimal in terms of coding gain for specificconditions.
P P Vaidyanathan stated that coding gain does not necessarily always increase with the increase inthe number of bands. However, very few attempts are made
in the literature to go beyond the confines of work done by P P Vaidyanathan. We present analytic proofs for the monotonicity of specific shapes of PSDs.
This papers also derives properties of coding gain of PCFBs, which brings the new insights on the coding gain of Principal Component Filter Banks.</i></p>
</td>
</tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/qr_hopfield_denoise.PNG" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>QR Code Denoising using Parallel Hopfield Networks</h3>
<br>
<strong>Shubhang Bhatnagar</strong><sup>*</sup>, Ishan Bhatnagar<sup>*</sup>
<br><strong>Arxiv</strong> , 2018
<br>
<a href="javascript:toggleblock('qr_abs')">abstract</a> /
<a href="https://arxiv.org/abs/1812.01065">arxiv</a>
<br>
<p align="justify"> <i id="qr_abs">We propose a novel algorithm for using Hopfield
networks to denoise QR codes. Hopfield networks have mostly been
used as a noise tolerant memory or to solve difficult combinatorial
problems. One of the major drawbacks in their use in noise tolerant
associative memory is their low capacity of storage, scaling only
linearly with the number of nodes in the network. A larger capacity
therefore requires a larger number of nodes, thereby reducing the
speed of convergence of the network in addition to increasing
hardware costs for acquiring more precise data to be fed to a larger
number of nodes. Our paper proposes a new algorithm to allow the
use of several Hopfield networks in parallel thereby increasing the
cumulative storage capacity of the system many times as compared
to a single Hopfield network. Our algorithm would also be much
faster than a larger single Hopfield network with the same total
capacity. This enables their use in applications like denoising QR
codes, which we have demonstrated in our paper. We then test our
network on a large set of QR code images with different types of
noise and demonstrate that such a system of Hopfield networks can
be used to denoise and recognize QR codes in real time.</i></p>
</td>
</tr>
<td style="padding:2.5%;width:40%;vertical-align:top;min-width:120px">
<img src="images/adapter_img.PNG" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:60%;vertical-align:top">
<h3>Memory Efficient Adaptive Attention For Multiple Domain Learning</h3>
<br>
Himanshu Pradeep Aswani, Abhiraj Sunil Kanse, <strong>Shubhang Bhatnagar</strong>, <a href="https://www.ee.iitb.ac.in/~asethi/">Amit Sethi</a>
<br><strong>Arxiv</strong> , 2021
<br>
<a href="javascript:toggleblock('adapter_abs')">abstract</a> /
<a href="https://arxiv.org/abs/2110.10969">arxiv</a>
<br>
<p align="justify"> <i id="adapter_abs">We propose a novel algorithm for using Hopfield
networks to denoise QR codes. Hopfield networks have mostly been
used as a noise tolerant memory or to solve difficult combinatorial
problems. One of the major drawbacks in their use in noise tolerant
associative memory is their low capacity of storage, scaling only
linearly with the number of nodes in the network. A larger capacity
therefore requires a larger number of nodes, thereby reducing the
speed of convergence of the network in addition to increasing
hardware costs for acquiring more precise data to be fed to a larger
number of nodes. Our paper proposes a new algorithm to allow the
use of several Hopfield networks in parallel thereby increasing the
cumulative storage capacity of the system many times as compared
to a single Hopfield network. Our algorithm would also be much
faster than a larger single Hopfield network with the same total
capacity. This enables their use in applications like denoising QR
codes, which we have demonstrated in our paper. We then test our
network on a large set of QR code images with different types of
noise and demonstrate that such a system of Hopfield networks can
be used to denoise and recognize QR codes in real time.</i></p>
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td><br>
<p align="right">
<font size="2">
Template: <a href="https://jonbarron.info">this</a>
</font>
</p>
</td>
</tr>
</table>
<script xml:space="preserve" language="JavaScript">
hideblock('potential_field_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('MLR2_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('MLR1_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('PL_UDML_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('ldgest_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('l1_cv_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('pal_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('pcfb_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('qr_abs');
</script>
<script xml:space="preserve" language="JavaScript">
hideblock('adapter_abs');
</script>
<br>
<br>
<br>
<!--
<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">
<h2>Intel Projects</h2>
<p>Besides my work on the RealSense depth sensors and the publications above, a sampling of my publicly disclosed work
</p>
</td>
</tr>
</table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
{% for post in site.posts %}
{% for cat in post.categories %}
{% if cat == 'Intel' %}
<tr>
<td style="padding:2.5%;width:25%;vertical-align:middle;min-width:120px">
<img src="/tn{{post.image}}" alt="project image" style="width:auto; height:auto; max-width:100%;" />
</td>
<td style="padding:2.5%;width:75%;vertical-align:middle">
<h3>{{post.title}}</h3>
<br>
<em>{{post.categories}} {{post.course}}</em>
<br>
{{ post.date | date: "%Y-%m-%d" }}
<br>
{% if post.website %}
<a href="{{post.website}}">website</a> /
{% endif %}
{% if post.paper %}
<a href="{{post.paper}}">paper</a> /
{% endif %}
{% if post.patent %}
<a href="{{post.patent}}">patent</a> /
{% endif %}
{% if post.patent2 %}
<a href="{{post.patent2}}">patent #2</a> /
{% endif %}
{% if post.patent3 %}
<a href="{{post.patent3}}">patent #3</a> /
{% endif %}
{% if post.video %}
<a href="{{post.video}}">video</a> /
{% endif %}
{% if post.video2 %}
<a href="{{post.video2}}">video #2</a> /
{% endif %}
{% if post.code %}
<a href="{{post.code}}">code</a> /
{% endif %}
{% if post.poster %}
<a href="{{post.poster}}">poster</a> /
{% endif %}
{% if post.slides %}
<a href="{{post.slides}}">slides</a> /
{% endif %}
<p></p>
{{ post.excerpt }}
</td>
</tr>
{% endif %}
{% endfor %}
{% endfor %}
</table>-->
</body>
</html>