-
Notifications
You must be signed in to change notification settings - Fork 0
/
About.html
710 lines (586 loc) · 33.7 KB
/
About.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
<!-- ---
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 rgb(138, 129, 129);
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;
max-height:100%;
max-width:100%;
/*margin-bottom: -10px;*/
}
.img_container {
position: relative;
width: 100%;
padding-bottom: 80%;
float:left;
}
.image_c {
position:absolute;
width:100%;
height:100%;
}
.image_c img {
position: absolute;
top: 0;
bottom: 0;
left: 0;
right: 0;
margin: auto;
}
.zero-padding {
padding: 0px !important;
}
</style>
</head>
<body>
<!--
<section class="container" style="width:80%;">
<nav class="site-nav">
<h1>
<a class="page-link" href="/">Bansal Blog!!</a>
</h1>
</nav>
<blockquote class="blockquote bg-faded">
<img class="img-fluid" align="right" src="assets/images/profile_new.jpg" width="200" height="240" padding-left="20px"/>
<div class="row" style="border:none;">
<div class="col-xs-11 container">
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>
</div>
</div>
</blockquote>
</section> -->
<section class="container" style="width:80%;">
<nav class="site-nav">
<h1>
<a class="page-link" href="/">Bansal Blog!!</a>
</h1>
</nav>
<div class="row" style="border:none;">
<div class="col-sm-4 col-xs-4">
<br>
<img class="img-fluid" align="right" src="assets/images/profile_new.jpg" width="240" height="320" padding-left="20px"/>
</div>
<div class="col-sm-7 col-xs-8">
Myself Ankish Bansal. I completed my M.Tech Degree in Jan, 2020 under the supervision of Prof.<a href="http://home.iitk.ac.in/~lbehera/" target="_blank"> L. Behera</a> at <a href="http://www.iitk.ac.in/" target="_blank">IIT, Kanpur</a>. My master thesis was on Meta Learning and Unsupervised Learning methods with the applications in Vision and Robotics.
<!-- 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 other interest areas include <strong>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>
</div>
</div>
</section>
<!-- <div class="container">
<div class="row">
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece">
<div class="img_container">
<div class="image_c">
<img class="img-fluid" align="right" src="assets/images/profile_new.jpg" width="240" height="320" padding-left="20px"/>
</div>
</div>
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece">
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>
</div>
</div>
</div>
</div> -->
<br>
<br>
<br>
<div class="container" style="width: 85%">
<h2>Following is the glossary of my data-science skills.</h2>
</div>
<br>
<!-- <section class="container" style="width:80%;">
<h1 align="center">
<a href="https://github.com/ankishb/ml-toolbox" target="_blank">ML Toolbox</a>
</h1>
<div class="container">
<img class="img-fluid" align="left" src="assets/images1/data-science.jpg" width="400" height="320" padding-left="20px"/>
<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>
<br>
</div>
</section> -->
<div class="container" style="width: 85%">
<div class="row">
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece">
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/Object-detection.jpg">
</div>
</div>
<!-- style="color:red;" -->
<h3 align="middle" >
<a href="https://github.com/ankishb/ml-projects/tree/master/conditional-object-detection" target="_blank">Class Agnostic Object Detection</a>
</h3>
<p>This is a Class Agnostic Detection problem, where we need to localize the object irrespective of the class.. My approach is similar to single shot detector algorithm such as YOLO and SSD, except the feature extraction pipeline. To extract class agnostic feature, I used conditional feature to put more attention on the object. This trick helped me to secure a rank of <strong>36/6733</strong> in the competition at dare2complete platform, sponsored by Flipkart.</p>
<small>Python, Tensorflow, Augmentation, YOLO, SSD</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/text-review.png">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/amazon-ml" target="_blank">Text classification</a>
</h3>
<p>The objective of this project is to design a model for <strong>Amazon Product Review Classification</strong>. This was Multilabel Classification Problem. My approach includes feature preprocessing, engineering and finally built an ensemble of model such as SVM, Logistic Regression, Decision Tree, Attention Model and BERT to accurately categorize the Amazon Product Review. This ensemble helped me to land in top shortlist for Amazon-Business-Intelligence profile.</p>
<!-- <a href="https://github.com/ankishb/ml-projects/tree/master/amazon-ml" target="_blank" class="btn btn-primary">Project</a> -->
<small>EDA, Scikit-Learn, Xgboost, Embedding, BERT, Ensemble</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/credit-risk.jpg">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/hdfc-ml" target="_blank">HDFC Risk Prediction</a>
</h3>
<p>This is my one of favorite project. The challege was to design a model for Risk Prediction for a financial coorporation (HDFC). We were given approximately 2500 unknown predictors. I experimented with several feature selection and feature engineering methods, during the contest. And finally, designed an efficient algorithm for interaction based feature, along with feature selection using decision tree. This repo also ha report on broad analysis of these predictors.</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, GBM, H2o, Stack-Net</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/social-network.jpg">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/hike-friend-recommendation" target="_blank">Hike Friend Recommendation</a>
</h3>
<p>This is a link prediction challenge for <strong>Hike Messenger</strong>. We were given a very big dataset, which doesn't fit in memory. During this project, I worked on feature engineering for graph dataset, graph neural network, memory optimization etc. My final approach is to collect all the feature and run LightGBM and CatBoost. My model secured a rank of <strong>32/5389</strong> in ML-Hikeathon contest. After the competition, I worked on subsemble model for large dataset.</p>
<!-- <a href="https://github.com/ankishb/ml-projects/tree/master/hike-friend-recommendation" target="_blank" class="btn btn-primary">Project</a> -->
<small>GraphX, GBM, Keras, Word-Embedding, Subsemble</small>
</div>
<div class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/image-classification.jpg">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/cifar-10-resnet" target="_blank">Cifar-10 Classification</a>
</h3>
<p>This is my research project on deep learning. The objective is to undertand the image feature used by deep cpnvolution feature and improve the model accuracy on Cifar-10 dataset. I designed an architecture, where base model is ResNet, but head model extract feature by using global as well as local feature of the image. This technique improves accuracy by <strong>1.37%</strong>. As the designed method igores the irrelevant information in the input image, therefore final attention map is much better than simple ResNet</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-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/facenet.jpg">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/facenet" target="_blank">Face Verification system</a>
</h3>
<p>This was my college project, where objective was to build a student-attendence system. Its input is image of a person and output is a verification task, if person is registered in the course. The collected dataset was very small. So to train a deep neural network model, I adopted <strong>network in network</strong> architectureI for my <strong>Maching Network</strong>. I trained it using hard-mining technique, and got <strong>93%</strong> accuraacy.</p>
<!-- <br> -->
<!-- <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-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/segmentation1.jpg">
</div>
</div>
<h3 align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/segmentation" target="_blank">Segmentation</a>
</h3>
<p>In this project, I first implemented an <strong>U-Net</strong> architecture and trained on blood cell Dataset (on Kaggle). I extend it to multi-stage network, to improve score. I also experimented with differnt configuration of <strong>Fully Convolutional Network</strong> for traffic-street dataset. Finally I experimented with <strong>Generative Adverserial Network</strong> for Data Augmentation, to learn better feature (But sadly did not get success).</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-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/recommender-sytem.jpg">
</div>
</div>
<h3 class="box-title" align="middle">
<a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system" target="_blank">Recommender System</a>
</h3>
<p>The objective of this problem is to recommend question to users, on a coding platform. So we have to answer question like, which question user should solve next? How about when user get stuck? etc. My approach is a hybrid approach using matrix factorization, content based, collaborative based and deep and shallow model. I used TF-IDF and word embedding, both for feature extraction in my model.</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 class="col-sm-6 col-md-4 col-lg-4" style="border:2px solid #cecece;" >
<div class="img_container">
<div class="image_c">
<img align="middle" width="400" height="240" src="assets/images1/data-science.jpg">
</div>
</div>
<h3 class="box-title" align="middle">
<a href="https://github.com/ankishb/ml-toolbox" target="_blank">ML Toolbox</a>
</h3>
<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, by following the Research Paper and advices on kaggle.</p>
<!-- <a href="https://github.com/ankishb/ml-projects/tree/master/recommendation-system" target="_blank" class="btn btn-primary">Project</a> -->
<small>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><br>
<div class="container" style="width: 80%">
<h3>Small Fun Projects</h3>
<div class="container">
<div class="container">
<h4><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/gartner">Gartner Retention Status Prediction</a></h4>
</div>
<div class="container" style="width: 98%">
<ul>
<li>This was one of first project on real world raw dataset</li>
<li>Learnt about usuage of Scikit Learn packages</li>
<li>Experiemnt with cross validation strategy</li>
<li>Used EDA and Feature Preprocessing, finallu built a Xgboost model</li>
</ul>
</div>
</div>
<div class="container">
<div class="container">
<h4><a href="https://github.com/ankishb/ml-projects/tree/master/ltfs-loan-prediction">LTFS Loan Status prediction</a></h4>
</div>
<div class="container" style="width: 98%">
<ul>
<li>Main challenge was the noise in dataset.</li>
<li>Extensive Feature Exploration using EDA tools</li>
<li>Feature Preprocessing such that transformation, binning etc</li>
<li>Used Auto-ML tool from H2o library</li>
<li>Final model was Ensemble of Gradient Boosting Models</li>
</ul>
</div>
</div>
<div class="container">
<div class="container">
<h4><a href="https://github.com/ankishb/ml-projects/tree/master/jp-morgan">JP.Morgan House Price Prediction</a></h4>
</div>
<div class="container" style="width: 98%">
<ul>
<li>This is a advance regression problem</li>
<li>As dataset was noisy, so base model was not doing good</li>
<li>I used StackNet approach, to design a model for meta feature</li>
</ul>
</div>
</div>
<div class="container">
<div class="container">
<h4><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/future-sale-pred">Future sale Prediction</a></h4>
</div>
<div class="container" style="width: 98%">
<ul>
<li>Objective is to predict sale of each store, based on history of products and store.</li>
<li>I learnt, how to handle time sequenced dataset.</li>
<li>Experiment with time involved feature engineering</li>
<li>Learnt about trend, season, time series predictive models such as AR, MA, ARMA and ARIMA</li>
</ul>
</div>
</div>
<div class="container">
<div class="container">
<h4><a href="https://github.com/ankishb/ml-projects/tree/master/small-fun-project/collect-imp-tensor-spyder/time-series-prediction">Stock Prediction</a></h4>
</div>
<div class="container" style="width: 98%">
<ul>
<li>Objective is to predict future value based on history</li>
<li>Used recurrent neural network</li>
<li>Experimented on mutli regression model</li>
</ul>
</div>
</div>
</div>
<br>
<br>
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js"></script>
</body>
</html>