-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
414 lines (379 loc) · 22.2 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
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta name=viewport content="width=device-width, initial-scale=1">
<meta name="generator" content="HTML Tidy for Linux/x86 (vers 11 February 2007), see www.w3.org">
<style type="text/css">
/* Color scheme stolen from Sergey Karayev */
a {
color: #1772d0;
text-decoration:none;
}
a:focus, a:hover {
color: #f09228;
text-decoration:none;
}
body,td,th,tr,p,a {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 14px
}
strong {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 14px;
}
heading {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 22px;
}
papertitle {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 14px;
font-weight: 700
}
name {
font-family: 'Lato', Verdana, Helvetica, sans-serif;
font-size: 32px;
}
.one
{
width: 160px;
height: 160px;
position: relative;
}
.two
{
width: 160px;
height: 160px;
position: absolute;
transition: opacity .2s ease-in-out;
-moz-transition: opacity .2s ease-in-out;
-webkit-transition: opacity .2s ease-in-out;
}
.fade {
transition: opacity .2s ease-in-out;
-moz-transition: opacity .2s ease-in-out;
-webkit-transition: opacity .2s ease-in-out;
}
span.highlight {
background-color: #ffffd0;
}
</style>
<title>Yash Goyal</title>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<link href='https://fonts.googleapis.com/css?family=Lato:400,700,400italic,700italic' rel='stylesheet' type='text/css'>
</head>
<body>
<table width="950" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="97%" valign="middle">
<p align="center">
<name>Yash Goyal</name>
<p align="center">Email: yashgoyal.yg1-at-gmail.com</p>
</p>
<p>
I am a Research Scientist at <a target="_blank" href="https://www.sait.samsung.co.kr/saithome/about/labs.do">Samsung - SAIT AI Lab Montreal</a> within <a target="_blank" href="https://mila.quebec/en">Mila</a>.
</p>
<p>
Previously, I was a PhD student in the <a target="_blank" href="https://www.ic.gatech.edu/">School of Interactive Computing</a> at <a target="_blank" href="http://www.gatech.edu/">Georgia Tech</a>. I was advised by <a target="_blank" href="https://www.cc.gatech.edu/~dbatra/">Dhruv Batra</a>. I also collaborated closely with <a target="_blank" href="https://www.cc.gatech.edu/~parikh/">Devi Parikh</a>.
</p>
<p> I used to co-organize the annual <a target="_blank" href="https://visualqa.org/">VQA Challenge</a>.</p>
<p> As a research intern, I have spent time at <a target="_blank" href="https://ai.google/research/teams/brain/">Google Brain</a> in spring+summer 2019,
<a target="_blank" href="https://research.fb.com/category/facebook-ai-research-fair/">Facebook AI Research</a> in spring+summer 2017, at Army Research Laboratory (ARL) Adelphi in summer 2015, and at Duke University in summer 2013.
<p align=center>
<a href="mailto:[email protected]">Email</a>  / 
<a target="_blank" href="YashGoyal-CV.pdf">CV</a>  / 
<a href="https://scholar.google.com/citations?user=Xa82XjQAAAAJ&hl=en">Google Scholar</a>  / 
<a href="https://arxiv.org/search/cs?searchtype=author&query=Goyal%2C+Y">arXiv</a>
</p>
</td>
<td >
<img src="yash.jpg" style="height:240px;">
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="100%" valign="middle">
<heading>Research</heading>
<p>Image Generative Models, Bias, Explainable AI, Vision & Language.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<heading>Publications</heading>
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td width="30%" align="center">
<img src='./images/mapl.png' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/2210.07179">
<papertitle>MAPL: Parameter-Efficient Adaptation of Unimodal Pre-Trained Models for Vision-Language Few-Shot Prompting</papertitle></a><br>
<a href="https://oscmansan.github.io/">Oscar Manas</a>, <a href="https://prlz77.github.io/">Pau Rodriguez*</a>, <a href="https://www.linkedin.com/in/saba-ahmadi">Saba Ahmadi*</a>, <a href="http://www.aidanematzadeh.me/">Aida Nematzadeh</a>, <strong>Yash Goyal</strong>, <a href="https://www.iro.umontreal.ca/~agrawal/">Aishwarya Agrawal</a> <br>
<p></p>
<em>The European Chapter of the Association for Computational Linguistics (EACL)</em>, 2023<br>
<br>
</td>
</tr>
<tr>
<td width="30%" align="center">
<img src='./images/imagecode.png' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/2203.15867">
<papertitle>Image Retrieval from Contextual Descriptions</papertitle></a><br>
Benno Krojer, Vaibhav Adlakha, Vibhav Vineet, <strong>Yash Goyal</strong>, Edoardo Ponti, Siva Reddy <br>
<p></p>
<em>The Association for Computational Linguistics (ACL)</em>, 2022<br>
<br>
</td>
</tr>
<tr>
<td width="30%" align="center">
<img src='./images/equas.png' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://onlinelibrary.wiley.com/doi/pdf/10.1002/ail2.60">
<papertitle>Reframing explanation as an interactive medium: The EQUAS (Explainable QUestion Answering System) project</papertitle></a><br>
William Ferguson, Dhruv Batra, Raymond Mooney, Devi Parikh, Antonio Torralba, David Bau, David Diller, Josh Fasching, Jaden Fiotto-Kaufman, <strong>Yash Goyal</strong>, Jeff Miller, Kerry Moffitt, Alex Montes de Oca, Ramprasaath R Selvaraju, Ayush Shrivastava, Jialin Wu, Stefan Lee<br>
<p></p>
<em>Applied AI Letters</em>, 2021<br>
<br>
</td>
</tr>
<tr>
<td width="30%" align="center">
<img src='./images/cace.png' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/1907.07165.pdf">
<papertitle>Explaining Classifiers with Causal Concept Effect (CaCE)</papertitle></a><br>
<strong>Yash Goyal</strong>, Amir Feder, <a href="https://web.iem.technion.ac.il/en/people/urishalit.html">Uri Shalit</a>, <a href="https://beenkim.github.io/">Been Kim</a><br>
<p></p>
<em>ArXiv</em>, 2019<br>
<br>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/counterfactual_explanations.png' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/1904.07451.pdf">
<papertitle>Counterfactual Visual Explanations</papertitle></a><br>
<strong>Yash Goyal</strong>, <a href="http://wuziyan.com/">Ziyan Wu</a>, <a href="https://www.linkedin.com/in/jan-ernst-2829965">Jan Ernst</a>,
<a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a>, <a href="https://computing.ece.vt.edu/~parikh/">Devi Parikh</a>, <a href="https://www.cc.gatech.edu/~slee3191/">Stefan Lee</a> <br>
<p></p>
<em>International Conference on Machine Learning (ICML)</em>, 2019<br>
<br>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/vqa_v2.jpg' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://link.springer.com/article/10.1007/s11263-018-1116-0">
<papertitle>Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering</papertitle></a><br>
<strong>Yash Goyal</strong>, <a href="https://tejaskhot.github.io/">Tejas Khot</a>, <a href="https://computing.ece.vt.edu/~aish/">Aishwarya Agrawal</a>, Douglas Summers-Stay,
<a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a>, <a href="https://computing.ece.vt.edu/~parikh/">Devi Parikh</a> <br>
<p></p>
<em>International Journal of Computer Vision (IJCV)</em>, 2018<br>
<a href="http://www.visualqa.org/">Project Website</a>,
<a href="https://www.youtube.com/watch?v=nMr_sSAMpkE">Demo </a>
<br>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/teaser_holistic.PNG' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="http://arxiv.org/pdf/1604.02125.pdf">
<papertitle>Resolving Language and Vision Ambiguities Together: Joint Segmentation & Prepositional Attachment Resolution in Captioned Scenes</papertitle></a><br>
<a href="https://computing.ece.vt.edu/~gordonac/">Gordon Christie*, Ankit Laddha*, <a href="https://computing.ece.vt.edu/~aish/">Aishwarya Agrawal</a>, <a href="https://computing.ece.vt.edu/~santol/">Stanislaw Antol</a>, <strong>Yash Goyal</strong>, <a href="http://www.me.vt.edu/people/faculty/kevin-kochersberger/">Kevin Kochersberger</a>, <a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a> <br>
*equal contribution<br>
<p></p>
<em>Computer Vision and Image Understanding (CVIU)</em>, 2017
<br>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/vqa_v2.jpg' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/1612.00837.pdf">
<papertitle>Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering</papertitle></a><br>
<strong>Yash Goyal</strong>*, <a href="https://tejaskhot.github.io/">Tejas Khot</a>*, Douglas Summers-Stay,
<a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a>, <a href="https://computing.ece.vt.edu/~parikh/">Devi Parikh</a> <br>
*equal contribution<br>
<p></p>
<em>Computer Vision and Pattern Recognition (CVPR)</em>, 2017<br>
<!-- <a href="https://arxiv.org/abs/1612.00837">arXiv</a>, -->
<a href="http://www.visualqa.org/">Project Website</a>,
<a href="https://www.youtube.com/watch?v=nMr_sSAMpkE">Demo </a>
<br>
<p></p>
<p>
We counter the language priors present in the popular Visual Question Answering (VQA) dataset (Antol et al., ICCV 2015) and make vision (the V in VQA) matter! Specifically, we balance the VQA dataset by collecting complementary images such that every question in our balanced dataset is associated with not just a single image, but rather a pair of similar images that result in two different answers to the question. Our dataset is by construction more balanced than the original VQA dataset and has approximately twice the number of image-question pairs. Our complete balanced dataset will be publicly released as part of the 2nd iteration of the Visual Question Answering Challenge (VQA v2.0).
<p></p>
</a></p>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/teaser_holistic.PNG' width="100%" height="auto" style="border-style: none">
</td>
<td valign="top" width="70%">
<p><a href="http://arxiv.org/pdf/1604.02125.pdf">
<papertitle>Resolving Language and Vision Ambiguities Together: Joint Segmentation & Prepositional Attachment Resolution in Captioned Scenes</papertitle></a><br>
Gordon Christie*, Ankit Laddha*, <a href="https://computing.ece.vt.edu/~aish/">Aishwarya Agrawal</a>, <a href="https://computing.ece.vt.edu/~santol/">Stanislaw Antol</a>, <strong>Yash Goyal</strong>, <a href="http://www.me.vt.edu/people/faculty/kevin-kochersberger/">Kevin Kochersberger</a>, <a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a> <br>
*equal contribution<br>
<p></p>
<em>Conference on Empirical Methods in Natural Language Processing (EMNLP)</em>, 2016
<br>
<p></p>
<p>
We present an approach to simultaneously perform semantic segmentation and prepositional phrase attachment resolution for captioned images. We show that our vision and language modules have complementary strengths, and that joint reasoning produces more accurate results than any module operating in isolation.
<p></p>
</a></p>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/vqa-interpretability.jpg' width="250" height="auto" style="border-style: none">
</td>
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/1608.08974.pdf">
<papertitle>Towards Transparent AI Systems: Interpreting Visual Question Answering Models</papertitle></a><br>
<strong>Yash Goyal</strong>, <a href="https://computing.ece.vt.edu/~akrit/">Akrit Mohapatra</a>,
<a href="https://computing.ece.vt.edu/~parikh/">Devi Parikh</a>, <a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a><br>
<p></p>
<em>International Conference on Machine Learning (ICML) Workshop on Visualization for Deep Learning</em>, 2016 <br>
<font color="green">[Best Student Paper]</font> <br>
Interactive Visualizations: <a href="https://mlp.ece.vt.edu/masked_ques_vis/"> Question </a> and <a href="https://mlp.ece.vt.edu/masked_image_vis/"> Image </a>
<br>
<p></p>
<p>
In this paper, we experimented with two visualization methods -- guided backpropagation and occlusion -- to interpret deep learning models for the task of Visual Question Answering. Specifically, we find what part of the input (pixels in images or words in questions) the VQA model focuses on while answering a question about an image.
<p></p>
</a></p>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/binaryVQA.jpg' width="250" height="auto" style="border-style: none">
</td>
</td>
<td valign="top" width="70%">
<p><a href="https://arxiv.org/pdf/1511.05099.pdf">
<papertitle>Yin and Yang: Balancing and Answering Binary Visual Questions</papertitle></a><br>
<a href="https://filebox.ece.vt.edu/~penghuiwin/">Peng Zhang</a>*, <strong>Yash Goyal</strong>*,
Douglas Summers-Stay, <a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a>, <a href="https://computing.ece.vt.edu/~parikh/">Devi Parikh</a><br>
*equal contribution<br>
<p></p>
<em>Computer Vision and Pattern Recognition (CVPR)</em>, 2016<br>
<a href="https://github.com/VT-vision-lab/abstract_binary_VQA"> Data and Code </a>
<br>
<p></p>
<p>
We balance the existing VQA dataset so that VQA models are forced to understand the image to improve their performance. We propose an approach that focuses heavily on vision and answers the question by visual verification. Dataset and Code will be available soon!
<p></p>
</a></p>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/Overview.png' width="100%" height="auto" style="border-style: none">
</td>
</td>
<td valign="top" width="70%">
<p><a href="https://mlp.ece.vt.edu/papers/2015/CloudCV/CloudCV.pdf">
<papertitle>CloudCV: Large-Scale Distributed Computer Vision as a Cloud Service</papertitle></a><br>
<a href="http://dexter1691.github.io/">Harsh Agrawal</a>, Clint Solomon Mathialagan, <strong>Yash Goyal</strong>, Neelima Chavali, Prakriti Banik, Akrit Mohapatra, Ahmed Osman, <a href="https://computing.ece.vt.edu/~dbatra/">Dhruv Batra</a> <br>
<p></p>
Book Chapter, <em>Mobile Cloud Visual Media Computing</em> <br>
Editors: Gang Hua, Xian-Sheng Hua. Springer, 2015.
<br>
<a href="http://cloudcv.org/">Website</a>
<p></p>
<p>
We present a comprehensive system to provide access to state-of-the-art distributed computer vision algorithms as a cloud service through a Web Interface and APIs.
<p></p>
</a></p>
</td>
</tr>
<tr>
<td width="30%">
<img src='./images/roman.PNG' width="100%" height="auto" style="border-style: none">
</td>
</td>
<td valign="top" width="70%">
<p><a href="http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arnumber=6926239">
<papertitle>Design of a physiologically informed virtual reality based interactive platform for individuals with upper limb impairment</papertitle></a><br>
Deepesh Kumar, <strong>Yash Goyal</strong>, Sunil Nair, Arvind Chauhan, <a href="http://www.iitgn.ac.in/faculty/electrical/uttama.htm">Uttama Lahiri</a> <br>
<p></p>
<em>IEEE International System on Robot and Human Interactive Communication (Ro-MAN 2014)</em>, UK
<br>
<p></p>
<p>
<p></p>
</a></p>
</td>
</tr>
</table>
<table width="100%" align="center" border="0" cellspacing="0" cellpadding="20">
<tr>
<td>
<heading>Talks</heading>
</td>
</tr>
</table>
<ul>
<li> <a href="https://slideslive.com/38916900/robust-statistics-and-interpretability">
"Counterfactual Visual Explanations"</a> (1:04:00 - 1:08:25) [<a href="./slides/icml19_counterfactual_visual_explanations.pdf">slides</a>] <br>
International Coference on Machine Learning (ICML), 2019.<br><br>
<br><br>
</li>
<li> <a href="https://www.youtube.com/watch?v=gz2VoDrvX-A&feature=youtu.be&t=1h06m29s">
"The Visual Question Answering Challenge"</a> (1:06:29 - 1:11:40) [<a href="./slides/vqa_challenge_cvpr18_competition_session.pdf">slides</a>] <br>
Special Session: Workshop Competitions, CVPR, 2018.<br><br>
<iframe width="280" height="160" src="https://www.youtube.com/embed/gz2VoDrvX-A?rel=0&start=3989" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
<br><br>
</li>
<li> <a href="https://www.youtube.com/watch?v=6SFJd3x-Nl8">
"Visual Question Answering Challenge 2018: Overview of Challenge, Winner Announcements, and Analysis of Results" [<a href="http://visualqa.org/static/slides/2018_workshop_yash_slides.pdf">slides</a>] <br>
<a href="http://www.visualqa.org/workshop.html">VQA Challenge and Visual Dialog Workshop</a>, CVPR, 2018.<br><br>
<iframe width="280" height="160" src="https://www.youtube.com/embed/6SFJd3x-Nl8" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
</li>
<li> <a href="https://www.youtube.com/watch?v=goEe4gx1yZg&list=PL_bDvITUYucCbIESoPz1qCbRhP99Ga6uc&index=2&t=38m38s">
"Visual Question Answering Challenge 2017: Overview of Dataset, Challenge, Winner Announcements, and Analysis of Results"</a> (38:38 - 47:12) [<a href="http://www.visualqa.org/static/slides/vqa_challenge_overview_2017.pptx">slides</a>] <br>
<a href="http://www.visualqa.org/workshop_2017.html">VQA Challenge Workshop</a>, CVPR, 2017.<br><br>
<iframe width="280" height="160" src="https://www.youtube.com/embed/goEe4gx1yZg?start=2318" frameborder="0" allow="autoplay; encrypted-media" allowfullscreen></iframe>
<br><br>
</li>
<li> <a href="http://techtalks.tv/talks/towards-transparent-visual-question-answering-systems/63026/">
"Towards Transparent Visual Question Answering Systems"</a> [<a href="./slides/icml16_visualization_workshop_interpretability_VQA.pptx">slides</a>] <br>
Visualization for Deep Learning Workshop, ICML, 2016.<br><br>
</li>
<li>
"Yin and Yang: Balancing and Answering Binary Visual Questions" [<a href="./slides/macv16_yinandyang.pptx">slides</a>] <br>
<a href="http://www.cs.jhu.edu/~areiter/JHU/macv2016.html">Mid-Atlantic Computer Vision (MACV) Workshop</a>, 2016.<br><br>
</li>
</ul>
<p></p>
<p></p>
<div align="right"><font size="2">[Website courtesy]: The website template is based on <a href="http://jonbarron.info/">Jon Barron</a>'s website.</font></div>
</td>
</tr>
</table>
</body>
</html>