-
Notifications
You must be signed in to change notification settings - Fork 5
Expand file tree
/
Copy pathml-projects.html
More file actions
407 lines (407 loc) · 31.3 KB
/
ml-projects.html
File metadata and controls
407 lines (407 loc) · 31.3 KB
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
<!doctype html>
<html lang="en" class="no-js">
<head>
<meta charset="utf-8">
<meta http-equiv="content-type" content="text/html;charset=utf-8" />
<meta property="og:locale" content="en-US">
<meta property="og:site_name" content="Abhinav's Webpage">
<meta property="og:title" content="Abhinav Bohra Machine Learning Projects">
<meta property="og:description" content="Abhinav Bohra Machine Learning Projects">
<meta property="og:url" content="https://abhinav-bohra.github.io/ml-projects.html">
<link rel="canonical" href="https://abhinav-bohra.github.io/ml-projects.html">
<link rel="icon" type="image/png" href="images/icon_transparent.png">
<meta name="HandheldFriendly" content="True">
<meta name="MobileOptimized" content="320">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta http-equiv="cleartype" content="on">
<meta name="description" content="Abhinav Bohra - Computer Science & Engineering, IIT Kharagpur, Work Experience - Machine Learning Projects">
<meta name="keywords" content="Abhinav Bohra, Computer Science, Engineering, IIT Kharagpur, Machine Learning Projects">
<meta name="author" content="Abhinav Bohra">
<!-- Google Analytics and MS Clarity -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-MNSPB7PT9W"></script>
<script src="assets/js/google_analytics.js" type="text/javascript"></script>
<script src="assets/js/microsoft_clarity.js" type="text/javascript"></script>
<script src="assets/js/jquery-3.7.0.min.js" type="text/javascript"></script>
<script src="assets/js/loader.js" type="text/javascript"></script>
<!-- CSS -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/bootstrap@3.3.7/dist/css/bootstrap.min.css" integrity="sha384-BVYiiSIFeK1dGmJRAkycuHAHRg32OmUcww7on3RYdg4Va+PmSTsz/K68vbdEjh4u" crossorigin="anonymous">
<link rel="stylesheet" href="assets/css/main.css">
<link rel="stylesheet" href="assets/css/main.bundle.css">
<title>Abhinav Bohra | ML Projects</title>
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "Abhinav Bohra",
"url": "https://abhinav-bohra.github.io/ml-projects.html",
"logo": "https://abhinav-bohra.github.io/images/jprofile.png",
"description": "Abhinav Bohra has made notable contributions to the field of machine learning, particularly in the areas of natural language processing and deep learning. His projects have focused on practical applications, including the development of ECTSum, a bullet point summarization system for earnings call transcripts, and CoeuSearch, an intelligent neural file search engine with semantic understanding capabilities. He has also explored the use of deep convolutional generative adversarial networks for video game level generation. Abhinav's work has received recognition, including a position on the SemEval 2022 Task 8 Leaderboard, showcasing his dedication and accomplishments in the field of machine learning.",
"address": {
"@type": "PostalAddress",
"addressLocality": "Ahmedabad",
"addressRegion": "Gujarat",
"postalCode": "380052",
"addressCountry": "India"
},
"contactPoint": {
"@type": "ContactPoint",
"telephone": " +91-97755-23111",
"email": "abhinavbohra01@gmail.com",
"contactType": "Personal Profile"
},
"sameAs": [
"https://www.facebook.com/abhinavbohra01",
"https://www.twitter.com/abhinavbohra01",
"https://www.linkedin.com/in/abhinav-bohra",
"https://github.com/abhinav-bohra",
"https://scholar.google.com/citations?user=F51Ct9oAAAAJ&hl=en"
]
}
</script>
</head>
<body>
<div id="header"></div>
<div id="main" role="main">
<div id="sidebar"></div>
<div class="archive">
<meta itemprop="headline" content="Abhinav Bohra - Machine Learning Projects">
<meta itemprop="description" content="Abhinav Bohra - Machine Learning Projects">
<h2 style="font-size: x-large;"><b>Machine Learning Projects</b></h2>
<!-- <div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/tacos.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">TACOS: A Tourist Review Dataset For Aspect Category Opinion Sentiment Prediction And Other ABSA Tasks</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Pawan Goyal, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2022-12">Dec 2022</time> - <time datetime="2023-6">Jun 2023</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;">Created TACOS, a new large scale review dataset in the tourism domain. Benchmarked the dataset using state-of-the-art summarization models such as Seq2Path, EMC-GCN and BMRC. Proposed FinBERT-T5 based paraphraser model with state-of-the-art results.<strong> Published Long Paper at <a href="https://2023.emnlp.org/" style="color:#52adc8">EMNLP 2023</a></strong><br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">CUDA</span>
<span class="tag">Django</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/Tourism-ACOS" style="color:#52adc8">TACOS: A Tourist Review Dataset For Aspect Category Opinion Sentiment Prediction And Other ABSA Tasks</a></p>
<p style="margin-top:-3%">Paper Link: <a href="https://arxiv.org/abs/2210.12467" style="color:#52adc8">Long Paper in EMNLP 2023 (Main Conference)</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/tacos_img1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div> -->
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/gs.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">ECTSum: Bullet Point Summarization of Long Earnings Call Transcripts</p>
<p class="subtitle is-6">In association with <strong>Goldman Sachs</strong> |
<time datetime="2022-4">Apr 2022</time> - <time datetime="2022-7">Jul 2022</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Created ECTSum, a new dataset using Earnings Call Transcripts (ECTs) of publicly traded companies as documents, and short expert-written summaries derived from corresponding Reuters articles. ECTs are long unstructured documents without any prescribed length limit or format. Benchmarked the dataset using state-of-the-art summarization models such as BigBird, SummaRuNNer and Longformer Encoder Decoder. Proposed FinBERT-T5 based paraphraser model with 13.3% ROUGE-2 gain and 8.5% less factual hallucination.<strong> Published Long Paper at <a href="https://2022.emnlp.org/" style="color:#52adc8">EMNLP 2022</a></strong><br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">CUDA</span>
</div>
<p>Project Link: <a href="https://github.com/rajdeep345/ECTSum" style="color:#52adc8">ECTSum: Bullet Point Summarization of Long ECTs</a></p>
<p style="margin-top:-3%">Paper Link: <a href="https://arxiv.org/abs/2210.12467" style="color:#52adc8">Long Paper in EMNLP 2022 (Main Conference)</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/gs_1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/vgl.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Video Game Level Generation using DCGAN</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Adway Mitra, Centre of Excellence in Artificial Intelligence, IIT Kharagpur |
<time datetime="2022-8">Aug 2022</time> - <time datetime="2022-11">Nov 2022</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Generating levels for video games using Machine Learning models instead of human designers is becoming increasingly common. In this paper, we explore an alternative GAN architecture applied to the creation of playable game levels with a focus on Super Mario games. We also compare latent space search techniques to optimise inputs to the GAN from within the latent vector space <br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">Java</span>
<span class="tag">Bash</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/VGL-GAN" style="color:#52adc8">Video Game Level Generation using DCGAN</a></p>
<p style="margin-top:-3%">Paper Link: <a href="https://github.com/abhinav-bohra/VGL-GAN/blob/main/Docs/VGL%20GAN%20Paper.pdf" style="color:#52adc8">VGL-GAN Paper</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/vgl_img1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/nfs.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Neural File Search Engine</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Palash Dey, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2021-8">Aug 2022</time> - <time datetime="2021-11">Nov 2022</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Designed and developed CoeuSearch, an NLP based intelligent local-file search engine that searches for relevant documents in a directory, considering the semantics of the file’s name as well as it's content. Invented three-fold search strategy using SBERT based dual encoders and KeyBERT Topic Extraction model. Employed cache optimization techniques to reduce response time by 70%<br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">NLTK</span>
<span class="tag">Django</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/CoeuSearch" style="color:#52adc8">Neural File Search Engine</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/nfs_1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/mna.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Multilingual News Article Similarity</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Pawan Goyal, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2022-1">Jan 2022</time> - <time datetime="2022-4">Apr 2022</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Leveraged the knowledge of pre-trained language models (mBERT and XLM) to predict the overall similarity between a given pair of articles. We propsed a model based on Sentence Transformer to estimate the contextualized embeddings coupled with cosine similarity. Our proposed approach using the Multilingual Setting is ranked 19th in the official SemEval 2022 Task 8 Leaderboard with a Pearson correlation score of 0.721. <br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">CUDA</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/NLP_Shared_Task_2022" style="color:#52adc8">Multilingual News Article Similarity</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/mna_img1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/efl.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Entailment as Few Shot Learner For ACOS Quad Extraction Task</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Pawan Goyal, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2021-12">Dec 2021</time> - <time datetime="2022-4">Apr 2022</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> In this work, we highlight limitations of generative models by doing extensive data analysis and present two novel approaches to address these limitations. One of them reformulates category classification into entailment task, while the other one uses paraphrase modeling paradigm to cast the ACOS task to a paraphrase generation process. Acknowledging the scarcity of specialized datasets across domains, we compare both in-domain & cross-domain performance of the considered methods for the ACOS task and report new state-of-the-art results.<br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">Paddle-NLP</span>
<span class="tag">CUDA</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/EFL-ACOS" style="color:#52adc8">Entailment as Few Shot Learner For ACOS Quad Extraction Task</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/efl_img1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/acos.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Investigating Generative Approaches For ACOS Quad Extraction Task</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Pawan Goyal, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2021-8">Aug 2021</time> - <time datetime="2021-11">Nov 2021</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Developed three generative methods for Aspect Category Opinion Sentiment (ACOS) task, two of which respect the order of generated triplets/quads by means of using autoregressive decoders, while the other leverages a novel set-based bipartite matching loss to train a non-autoregressive parallel decoder. Acknowledging the scarcity of specialized datasets across domains, compared both in-domain & cross-domain performance of the considered methods for the ASTE task, thereby drawing notable inferences. Employed all proposed architectures for the ACOS task and reported new state-of-the-art results on the corresponding benchmark dataset.<br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">PyTorch</span>
<span class="tag">Fast-AI</span>
<span class="tag">CUDA</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/Generative-Techniques-For-ACOS" style="color:#52adc8">Investigating Generative Approaches For ACOS Quad Extraction Tasks</a></p>
</div>
<div class="content col-lg-6" >
<img src="images/ml-projects/acos_img1.png" style="width:100%; height:285px; border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/emo.png" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Multitasking Framework for Emotional Analysis</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Pawan Goyal, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2021-1">Jan 2021</time> - <time datetime="2021-4">Apr 2021</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> This project is an implementation of the research paper <a href="https://www.cse.iitb.ac.in/~pb/papers/ieee-toac-sa.pdf" style="color:#52adc8">All-in-One: Emotion, Sentiment and Intensity Prediction using a Multi-task Ensemble Framework</a> which proposes a multi-task ensemble framework that jointly learns multiple related problems. The ensemble model aims to leverage the learned representations of three deep learning models (i.e., CNN, LSTM and GRU) and a hand-crafted feature representation for the predictions. Achieved 5.2% increase in accuracy and 0.33 increase in Pearson co-relation score for emotion classification and intensity tasks respectively.</p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">Keras</span>
<span class="tag">Tensorflow</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/Emotional-Analysis-Multitasking-Framework" style="color:#52adc8">Multitasking Framework for Emotional Analysis</a></p>
</div>
<div class="content col-lg-6">
<img class="mySlides" src="images/ml-projects/emo_img1.png" style="width:100%;height:285px;border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/spp.jpg" alt="Placeholder image" style="height: 100%;">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Stock Price Movement Prediction using Sentiment Analysis</p>
<p class="subtitle is-6"><strong>Advisor:</strong> Prof. Adway Mitra, Department of Computer Science & Engineering, IIT Kharagpur |
<time datetime="2021-8">Aug 2021</time> - <time datetime="2021-11">Nov 2021</time>
</p>
</div>
</div>
<div class="row">
<div class="content col-lg-6">
<p style="text-align: justify; text-justify: inter-word;"> Worked on establishing statistical correlation between social media sentiment and stock price movement of companies. Performed sentiment analysis using BERT on company's official tweets to generate social media sentiment score. Used it as an additional signal in LSTM network built on top of features like Open Stock price, Close Stock price, Low price, High price, Volume and Adj Close Price to predict the stock prices.</p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">Keras</span>
<span class="tag">Time Series</span>
<span class="tag">Sequence Models</span>
</div>
<p>Project Link: <a href="https://github.com/abhinav-bohra/Emotional-Analysis-Multitasking-Framework" style="color:#52adc8">Stock Price Movement Prediction using Sentiment Analysis</a></p>
</div>
<div class="content col-lg-6">
<img class="mySlides" src="images/ml-projects/spp_img1.png" style="width:100%;height:285px;border-radius: 10px;">
</div>
</div>
</div>
</div>
</div>
<div class="list__item" style="margin-top: 2%;">
<div class="card">
<div class="card-content">
<div class="media">
<div class="media-left">
<figure class="image is-48x48">
<img src="images/ml-projects/aurix.png" alt="Placeholder image">
</figure>
</div>
<div class="media-content">
<p class="title is-4">Aurix, Smart-Electroacoustic-Transducers</p>
<p class="subtitle is-6">Self Project |
<time datetime="2019-11">Nov 2019</time> - <time datetime="2022-11">Present</time>
</p>
</div>
</div>
<div class="content">
<p style="text-align: justify; text-justify: inter-word;">Imagine a world where you can immerse yourself in music while staying connected to the world around you. Enjoy your favorite tunes with your earphones or headphones, while still being alerted when someone is trying to reach you. For those times when you're on the move, this technology acts as your second set of ears, keeping you safe and alert. By alerting individuals using headphones while driving, we aim to significantly reduce the risk of accidents on the road.<br></p>
<div class="tags">
<span class="tag">Python</span>
<span class="tag">Speech Recognition</span>
<span class="tag">Natural Language Processing</span>
</div>
<!--p>Project Link: <a href="#" style="color:#52adc8">Aurix, Smart-Electroacoustic-Transducers</a></p-->
</div>
</div>
</div>
</div>
</div>
</div>
<div id="footer"></div>
</body>
</html>