-
Notifications
You must be signed in to change notification settings - Fork 0
/
About2.html
766 lines (632 loc) · 35.4 KB
/
About2.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
<!-- ---
layout: page
title: About
---
<img src="{{ '/assets/images/profile_new.jpg' | relative_url }}" width="240" height="320" align="right" />I
=======
<p> I'm writting about data scructure, deep learning and algorithms in ML, RL and in general. Send me an email: bansal.ankish1 AT gmail.com, if you have questions or find error in my posts. Thanks :)</p>
>>>>>>> 3d78cf1be9e114ad553ea2a9a9f26e38333dcd11
-->
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<meta http-equiv="x-ua-compatible" content="ie=edge">
<link rel="stylesheet" href="https://stackpath.bootstrapcdn.com/bootstrap/4.4.1/css/bootstrap.min.css" integrity="sha384-Vkoo8x4CGsO3+Hhxv8T/Q5PaXtkKtu6ug5TOeNV6gBiFeWPGFN9MuhOf23Q9Ifjh" crossorigin="anonymous">
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.4.1/css/bootstrap.min.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.4.1/jquery.min.js"></script>
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.4.1/js/bootstrap.min.js"></script>
<style>
.box {
position: relative;
width: 420px;
height: 580px;
border: 1px solid black;
box-sizing: border-box;
}
.box-s {
position: relative;
width: 900px;
height: 600px;
border: 1px solid black;
box-sizing: border-box;
}
.box-ss {
position: relative;
width: 900px;
height: 350px;
/*border: 1px solid blue;*/
box-sizing: border-box;
padding-left: 10px;
padding-right: 10px;
padding-top: 30px;
}
.box-text {
/*margin-top: -20px;*/
position: absolute;
width: 98%;
height: 50%;
margin-left: 2px;
margin-right: 2px;
}
.box-s-text {
/*margin-top: -20px;*/
width: 900px;
height: 180px;
margin-left: 10px;
margin-right: 10px;
}
.box-footer{
position: absolute;
bottom: 10px;
vertical-align: bottom;
display: inline-block;
margin-bottom: -10px;
width: 418px;
height: 60px;
background-color: lightgrey;
}
.box-s-footer{
text-align: center;
position: absolute;
font-size: 24px;
bottom: 10px;
vertical-align: bottom;
display: inline-block;
padding-top: 5px;
margin-bottom: -10px;
width: 898px;
height: 60px;
background-color: lightblue;
}
mark{
/*background-color: lightblue;*/
font-size: 14px;
/*color: black;*/
}
img{
border-radius: 10px;
}
</style>
</head>
<body>
<section class="container" style="width: 80%">
<nav class="site-nav">
<h1>
<a class="page-link" href="/">Bansal Blog!!</a>
</h1>
</nav>
<!-- <h1>About Me</h1> -->
<blockquote class="blockquote bg-faded">
<img class="img-fluid" align="right" src="assets/images/profile_new.jpg" width="240" height="320" padding-left="10px"/>My name is Ankish Bansal. I am a final year, M.Tech student at <a href="http://www.iitk.ac.in/" target="_blank">IIT, Kanpur</a>, where I am working in the field of Machine Learning under the supervision of Prof.<a href="http://home.iitk.ac.in/~lbehera/" target="_blank"> L. Behera</a>.
<br>
<br>
<p>My interest area includes <strong>Machine learning, Data Structure and Algorithms, Data Science, Database Management System and Statistical Learning</strong>. As a part of my M.Tech thesis, I work in Reinforcement Learning and Meta Learning. In Reinforcement Learning, the objective is to design techniques to deal with Credit Assignment problem, which is also known as Delayed Reward. In Meta Learning, the objective is fast adaptation and better generalization in very small dataset. It is similar to few-shot learning, multi-task learning and fine-tuning methods, except much powerful for adaptation over tasks.</p>
<br>
<p>Apart from my thesis work, I practice my coding skill (mostly in <strong>c++</strong> and <strong>pyhton</strong>), on coding platform such as hackerrank, interviewbit and leetcode. I also participate in ML-Hackhathon, with the objective to improve my data-science skill, work on different set of data and model, build practical insight of real world data.</p>
</blockquote>
</section>
<br>
<br>
<div class="container" style="width: 80%">
<h3>Glossary of my Data-Science Skills</h3>
<p>For each project, I described my approach in very short. More detail can be find in the repective repositary.</p>
</div>
<div class="container" style="width: 80%">
<section class="container">
<h3><a href="https://github.com/ankishb/ml-toolbox">ML Toolbox</a></h3>
<ul>
<li>This repo contains various data science strategy and machine learning models to deal with structure as well as unstructured data.</li>
<li>It contains module on feature-preprocessing, feature-engineering, machine-learning-models, etc.</li>
<li>Some of these features are collected from the existed libraries such as scikit-learn, keras, gensim, h2o, bayesopt, GraphX etc.</li>
<li>And others are implemented by me, on following the Research Paper and Data-Scientist advice (on kaggle).</li>
<li>I have worked on feature engineering strategy a lot, which can be find in this repo.</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection">Flipkart Object Detection</a></h3>
<ul>
<li>Designed a visual feature pipeline with attention on the object in image.
<li>Data Augmentation Technique along with its bounding box</li>
<li>Used Single Stage Detector Approach</li>
<li>Focal Loss with YOLO and SSD</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/blob/master/amazon-ml/submission-ankish-bansal/amazon_ml_final_submission.ipynb">Amazon Product Review classification</a></h3>
<ul>
<li>This is a multi-label classification task</li>
<li>Data Cleaning/feature enginnering</li>
<li>Linear/Non-Linear Model</li>
<li>Deep Learning Attention Model</li>
<li>Pretrained Bert Model</li>
<li>Ensemble</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/hdfc-ml">HDFC Risk Prediction</a></h3>
<ul>
<li>This is a classification problem</li>
<li>The data was very noisy, highly correlated, a lot of missing values, unknown types and outliers</li>
<li>Extensive Feature Preprocessing</li>
<li>Designed Feature Interaction algortihm</li>
<li>Write an efficient feature selection algorithm using recursive elimination method</li>
<li>ensemble model using xgboost/lighgbm/</li>catboost and linear/non-linear simple model
<li>statistical model to understand the feature importance using p-values</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/hike-friend-recommendation">Hike Friend Recommendation</a></h3>
<ul>
<li>Graph Based features such as (adamic-adar, common-resource-allocation,...)</li>
<li>SVD feature for each user</li>
<li>Comunity-clustering</li>
<li>Subsemble(I did this after competition is over, to understand more about sampling and model building)</li>
<li>neighbour-based feature(Removed highly cardinal feature)</li>
<li>Also tried Deep learning approach (Graph Embedding), but couldn't handle at that time properly</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/cifar-10-resnet">Cifar-10 Classification using Conditional Feature</a></h3>
<ul>
<li>Developed a weighted feature pipeline using global and local feature.</li>
<li>Global feature put constrained on local feature, to specifically focused on features of object in image</li>
<li>Better attention map around object, which reflect its learned feature.</li>
<li>Improved score by 1.37% over Resnet</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/facenet">FaceNet</a></h3>
<ul>
<li>Matching Network Approach</li>
<li>Build a Student-Attentdance hardware using arduino</li>
<li>Hard Mining Approac(generate all permutation between classes to handle small dataset)</li>
<li>network-in-network approach to handle overfitting as i have very small dataset.</li>
<li>Achieved 93% accuracy</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system">Hackerearth Platform Recommendation System</a></h3>
My main concerns was to handle following question carefully:
<ul>
<li>What is the strongest and weakest area of user?</li>
<li>What is the level of problem?</li>
<li>What problem user have just solved?</li>
<li>If user gets stuck at current problem, what problem should help him(to gain confidence and to improve skill in that area)?</li>
<li>Exploration and explotation strategy in recommending problem?</li>
<li>And many more?</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/segmentation">Segmentation</a></h3>
<ul>
<li>Implemented an U-Net architecture on blood cell Dataset.</li>
<li>fully convolutional network on traffic-street dataset.</li>
<li>Finally experimented with generative adverserial network for better generalization in the presence of limited dataset.</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/club-mahindra">Club Mahindra Hotel Room Price Prediction</a></h3>
<ul>
<li>Feature engineering</li>
<ul>
<li>date-time based feature</li>
<li>Aggregation based feature</li>
<li>Relational Features</li>
</ul>
<li>Ensemble using different set of tranformed target space</li>
</ul>
</section>
<section class="container">
<h3><a href="https://github.com/ankishb/ml-projects/tree/master/few-shot-classification">Few Shot Learning(Prototype Network)</a></h3>
<ul>
<li>Prototype Algorithm implementation.</li>
<li>There is more to this(will update in future)</li>
</ul>
</section>
</div>
<div class="container" style="width: 80%">
<h3>Small Fun Projects</h3>
<ul>
<li><a href="https://github.com/ankishb/ml-projects/tree/master/jp-morgan">JP.Morgan House Price Prediction</a></li>
<li><a href="https://github.com/ankishb/ml-projects/tree/master/ltfs-loan-prediction">LTFS Loan Status prediction</a></li>
<li><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/future-sale-pred">Future sale Prediction</a></li>
<li><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/gartner">Gartner Retention Status Prediction</a></li>
<li><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/collect-imp-tensor-spyder/time-series-prediction">Stock Prediction</a></li>
</ul>
</div>
<br><br><br>
<!-- <div class="container">
<img align="middle" height="240" src="assets/images1/data-science.jpg">
<h2 align="middle">ML Toolbox</h2>
<p>This repo contains various data science strategy and machine learning models to deal with structure as well as unstructured data. It contains module on feature-preprocessing, feature-engineering, machine-learning-models, etc. Some of these features are collected from the existed libraries such as scikit-learn, keras, gensim, h2o, bayesopt, xgboost, lightgbm, catboost, GraphX etc and others are implemented by me, on following the Research Paper and Data-Scientist advice (on kaggle). I have worked on feature engineering strategy a lot, which can be find in this repo.</p>
<a href="https://github.com/ankishb/ml-toolbox" target="_blank" class="btn btn-primary">Github Repo</a>
<br>
<small>Python, Gensim, Scikit-learn, H2o, bayesopt, GraphX, Keras, Xgboost, Catboost, lightgbm</small>
</div> -->
<!-- <br>
<br>
<br>
<br>
<br> -->
<!--
<section class="container text-center" id="services">
<div class="row">
<div class="col">
<img class="mx-auto" width="200" height="240" src="assets/images1/Object-detection.png">
<h3>Class Agnostic Object Detection</h3>
<p class="container">A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
<div class="container">
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
</div>
<article>
Python, Tensorflow, ImageAug, YOLO, SSD
</article>
</div>
<div class="col">
<img class="mx-auto" width="200" height="240" src="assets/images1/Object-detection.png">
<h3>Class Agnostic Object Detection</h3>
<p class="container">A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
<article class="container">
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
</article>
<article>
Python, Tensorflow, ImageAug, YOLO, SSD
</article>
</div>
</div>
<div class="row">
<div class="col">
<img class="mx-auto" width="200" height="240" src="assets/images1/Object-detection.png">
<h3>Class Agnostic Object Detection</h3>
<p class="container">A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
<article class="container">
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
</article>
<article>
Python, Tensorflow, ImageAug, YOLO, SSD
</article>
</div>
<div class="col">
<img class="mx-auto" width="200" height="240" src="assets/images1/Object-detection.png">
<h3>Class Agnostic Object Detection</h3>
<p>A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
<article class="container">
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
</article>
<article>
Python, Tensorflow, ImageAug, YOLO, SSD
</article>
</div>
</div>
</div>
</section> -->
<!--
<br>
<br>
<br>
<br>
<br>
-->
<!-- <div class="container" id="services">
<div class="row">
<section class="col-sm-6 col-md-4">
<img src="images/image.png" alt="sample image">
<h4>Exotic Pets</h4>
<p>We offer <strong>specialized</strong> care for <em>reptiles, rodents, birds,</em> and other exotic pets.</p>
</section>
<section class="col-sm-6 col-md-4">
<img src="images/image.png" alt="sample image">
<h4>Grooming</h4>
<p>Our therapeutic <span class="font-weight-bold">grooming</span> treatments help battle fleas, allergic dermatitis, and other challenging skin conditions.</p>
</section>
<section class="col-sm-6 col-md-4">
<img src="images/image.png" alt="sample image">
<h4>General Health</h4>
<p>Wellness and senior exams, ultrasound, x-ray, and dental cleanings are just a few of our general health services.</p>
</section>
<section class="col-sm-6 col-md-4">
<img class="img-fluid" src="images/image.png" alt="sample image">
<h4>Nutrition</h4>
<p>Let our nutrition experts review your pet's diet and prescribe a custom nutrition plan for optimum health and disease prevention.</p>
</section>
<section class="col-sm-6 col-md-4">
<img src="images/image.png" alt="sample image">
<h4>Pest Control</h4>
<p>We offer the latest advances in safe and effective prevention and treatment of fleas, ticks, worms, heart worm, and other parasites.</p>
</section>
<section class="col-sm-6 col-md-4">
<img src="images/image.png" alt="sample image">
<h4>Vaccinations</h4>
<p>Our veterinarians are experienced in modern vaccination protocols that prevent many of the deadliest diseases in pets.</p>
</section>
</div>
</div> -->
<!-- <br>
<br>
<br>
<br>
<br> -->
<!-- <div class="container">
<div class="row">
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Class Agnostic Object Detection</h3>
<p>A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
<small>Python, Tensorflow, ImageAug, YOLO, SSD</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Text classification</h3>
<p>An ensemble of classifiers such as SVM, Logistic Regression, Decision Tree, Attention Model and BERT to accurately categorize the Amazon Product Review.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/amazon-ml" target="_blank" class="btn btn-primary">Project</a>
<small>Python, EDA, Scikit-Learn, Xgboost, Keras, Word-Embedding, BERT, Ensemble</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">HDFC Risk Prediction</h3>
<p>In this project, the challege was to build a model for 2500 unknown predictors(features). During this project, I experimented on a lot of feature selection and feature engineering strategy, to deal with such kind of situation. </p>
<a href="https://github.com/ankishb/ml-projects/tree/master/hdfc-ml" target="_blank" class="btn btn-primary">Project</a>
<small>Python, EDA, Scikit-Learn, statsmodels, Gradient-Boosting, H2o, Stack-Net</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Hike Friend Recommendation</h3>
<p>This is a link prediction challenge. Here challenge is to handle a very big dataset, which doesn't fit in memory. During this project, I worked on graph network based feature engineering, graph embedding, memory optimization etc and secured a rank of 32/5389 in ML-Hikeathon contest.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/hike-friend-recommendation" target="_blank" class="btn btn-primary">Project</a>
<small>GraphX, Gradient-Boost, Keras, Word-Embedding, Subsemble</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Cifar-10 Classification</h3>
<p>Developed a architecture using state of the art ResNet, with class conditional feature. This add more attention on the class specific feature using global as well as local features. It improve the ResNet model's score by 1.37% on Cifar-10 Dataset.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/cifar-10-resnet" target="_blank" class="btn btn-primary">Project</a>
<small>Python, Tensorflow, Attention Mechanism, Visulization</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Face Verification system</h3>
<p>The objective is to build a student-attendence system which takes image as input and verify if the person is registered for the course or not. As dataset was very limited, I used network in network architecture along with hard-mining technique to successfully train a matching network model, to achieve 93% accuracy.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/facenet" target="_blank" class="btn btn-primary">Project</a>
<small>Tensorflow, Keras, ImageAug</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 align="middle">Segmentation</h3>
<p>Implemented an U-Net architecture on blood cell Dataset and fully convolutional network on traffic-street dataset. Finally experimented with generative adverserial network for better generalization in the presence of limited dataset.</p>
<a href="https://github.com/ankishb/ml-projects/tree/master/segmentation" target="_blank" class="btn btn-primary">Project</a>
<small>Python, Tensorflow, Keras, FCN, U-Net, GAN</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-3">
<h3 class="box-title" align="middle">Recommender System</h3>
<p class="box-text"></p>
<a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system" target="_blank" class="btn btn-primary">Project</a>
<small class="box-muted">Python, Tensorflow, Word-Embedding, Graph, Gradient-Boosting, Ensemble</small>
</div>
</div>
</div> -->
<!-- <div class="container">
<h2>My Data science Projects</h2>
<div class="container">
<table class="table">
<tbody>
<tr>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/Object-detection.png"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Class Agnostic Object Detection</h3>
<p class="box-text">A single shot detector algorithm in addition to object conditional feature, to create a robust bounding box around object. Using this method, I secured a rank of 36/6733 in the competition at dare2complete platform, sponsored by Flipkart.</p>
</div>
<div class="card-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Github Repo</a>
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank" class="btn btn-primary">Project Report</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, ImageAug, YOLO, SSD</small>
</div>
</div>
</div>
</td>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/text-review.png"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Text classification</h3>
<p class="box-text">An ensemble of classifiers such as SVM, Logistic Regression, Decision Tree, Attention Model and BERT to accurately categorize the Amazon Product Review.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/amazon-ml" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, EDA, Scikit-Learn, Xgboost, Keras, Word-Embedding, BERT, Ensemble</small>
</div>
</div>
</div>
</td>
</tr>
<tr><td><br></td></tr>
<tr>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/credit-risk.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">HDFC Risk Prediction</h3>
<p class="box-text">In this project, the challege was to build a model for 2500 unknown predictors(features). During this project, I experimented on a lot of feature selection and feature engineering strategy, to deal with such kind of situation. </p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/hdfc-ml" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, EDA, Scikit-Learn, statsmodels, Gradient-Boosting, H2o, Stack-Net</small>
</div>
</div>
</div>
</td>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/social-network.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Hike Friend Recommendation</h3>
<p class="box-text">This is a link prediction challenge. Here challenge is to handle a very big dataset, which doesn't fit in memory. During this project, I worked on graph network based feature engineering, graph embedding, memory optimization etc and secured a rank of 32/5389 in ML-Hikeathon contest.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/hike-friend-recommendation" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">GraphX, Gradient-Boost, Keras, Word-Embedding, Subsemble</small>
</div>
</div>
</div>
</td>
</tr>
<tr><td><br></td></tr>
<tr>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/image-classification.png"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Cifar-10 Classification</h3>
<p class="box-text">Developed a architecture using state of the art ResNet, with class conditional feature. This add more attention on the class specific feature using global as well as local features. It improve the ResNet model's score by 1.37% on Cifar-10 Dataset.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/cifar-10-resnet" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, Attention Mechanism, Visulization</small>
</div>
</div>
</div>
</td>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/facenet.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Face Verification system</h3>
<p class="box-text">The objective is to build a student-attendence system which takes image as input and verify if the person is registered for the course or not. As dataset was very limited, I used network in network architecture along with hard-mining technique to successfully train a matching network model, to achieve 93% accuracy.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/facenet" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Tensorflow, Keras, ImageAug</small>
</div>
</div>
</div>
</td>
</tr>
<tr><td><br></td></tr>
<tr>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/segmentation1.png"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Segmentation</h3>
<p class="box-text">Implemented an U-Net architecture on blood cell Dataset and fully convolutional network on traffic-street dataset. Finally experimented with generative adverserial network for better generalization in the presence of limited dataset.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/segmentation" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, Keras, FCN, U-Net, GAN</small>
</div>
</div>
</div>
</td>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/recommender-sytem.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Recommender System</h3>
<p class="box-text"></p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, Word-Embedding, Graph, Gradient-Boosting, Ensemble</small>
</div>
</div>
</div>
</td>
</tr>
<tr><td><br></td></tr>
<tr>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/stock-market.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Segmentation</h3>
<p class="box-text">Implemented an U-Net architecture on blood cell Dataset and fully convolutional network on traffic-street dataset. Finally experimented with generative adverserial network for better generalization in the presence of limited dataset.</p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/segmentation" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, Keras, FCN, U-Net, GAN</small>
</div>
</div>
</div>
</td>
<td>
<div>
<div class="box">
<center><img align="middle" width="400" height="240" src="assets/images1/text-mining.jpg"></center>
<div class="box-block">
<h3 class="box-title" align="middle">Recommender System</h3>
<p class="box-text"></p>
</div>
<div class="box-buttons">
<a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system" target="_blank" class="btn btn-primary">Project</a>
</div>
<div class="box-footer">
<small class="box-muted">Python, Tensorflow, Word-Embedding, Graph, Gradient-Boosting, Ensemble</small>
</div>
</div>
</div>
</td>
</tr>
</tbody>
</table>
</div> -->
<!-- <br>
<h1>Small Fun Projects</h1>
<ul>
<li class="list-group-item">
<div class="d-flex w-100 justify-content-between">
<h5>Behavioral Risk Factor Surveillance System 2013: Exploratory Data Analysis</h5>
<a target="_blank" href="/portfolio/brfss2013_eda">Notebook</a>
</div>
<p>Analysis of the BRFSS-2013 data set using R, focusing on investigating the relationship between education and eating habits, sleep and mental health, and smoking, drinking and general health of a person.</p>
<p class="list-group-skills">Skills: R, Descriptive Statistics, ggplot, dplyr</p>
</li>
<li class="list-group-item">
<div class="d-flex w-100 justify-content-between">
<h5>3-way Sentiment Analysis for Tweets</h5>
<a target="_blank" href="/portfolio/sentiment_analysis_tweets">Notebook</a>
</div>
<p>3-way polarity (positive, negative, neutral) classification system for tweets, without using NLTK's sentiment analysis engine.</p>
<p class="list-group-skills">Skills: Python, NLP, Scikit-learn</p>
</li>
<li class="list-group-item">
<div class="d-flex w-100 justify-content-between">
<h5>Inferential Statistics: Do men or women oppose sex education?</h5>
<a target="_blank" href="/portfolio/gss_inferential">Notebook</a>
</div>
<p>Analysing the GSS (General Social Survey) dataset using R to infer if, in the year 2012, were men, of 18 years or above in the United States, more likely to oppose sex education in public schools than women.</p>
<p class="list-group-skills">Skills: R, Hypothesis Testing, ggplot, dplyr</p>
</li>
</ul> -->
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js"></script>
</body>
</html>