-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
executable file
·377 lines (347 loc) · 19.3 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
<!DOCTYPE HTML>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Subash Khanal</title>
<meta name="author" content="Subash Khanal">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="shortcut icon" href="images/favicon.ico" type="image/x-icon">
<link rel="stylesheet" type="text/css" href="stylesheet.css">
</head>
<body>
<table style="width:100%;max-width:800px;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:0px">
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr style="padding:0px">
<td style="padding:2.5%;width:63%;vertical-align:middle">
<p class="name" style="text-align: center;">
Subash Khanal
</p>
<p>I am a PhD candidate in Computer Science at the <a href="https://wustl.edu/">Washington University in St. Louis </a>, working in <a href="https://mvrl.cse.wustl.edu/">Multimodal Vision Research Laboratory</a> led by <a href="https://engineering.wustl.edu/faculty/Nathan-Jacobs.html">Dr. Nathan Jacobs</a>.
</p>
<p>
I have a MS in Electrical Engineering from the <a href="https://www.uky.edu/">University of Kentucky</a> in <a href="https://en.wikipedia.org/wiki/Lexington,_Kentucky"> Lexington</a>.
During my masters, I worked with <a href="http://johnson.engr.uky.edu">Dr. Michael T. Johnson</a> focusing on speech recognition and signal processing.
</p>
<p style="text-align:center">
<a href="mailto:[email protected]">Email</a> /
<a href="data/Subash_Khanal_Resume_latest.pdf">CV</a> /
<a href="https://scholar.google.com/citations?user=G72g3R0AAAAJ&hl=en">Google Scholar</a> /
<a href="https://www.linkedin.com/in/subash-khanal/">Linkedin</a> /
<a href="https://github.com/subash-khanal/">Github</a>
</p>
</td>
<td style="padding:2.5%;width:40%;max-width:40%">
<a href="images/photo2_circular.jpg"><img style="width:100%;max-width:100%" alt="profile photo" src="images/photo2_circular.jpg" class="hoverZoomLink"></a>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<h2>Research</h2>
<p>
My research focuses on developing methods for Geospatial AI using Deep Learning and Computer Vision. Currently, I am working on multimodal self-supervised learning techniques to integrate satellite imagery, audio, and text for global-scale soundscape mapping.
</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:20px;width:100%;vertical-align:middle">
<h2>Publications</h2>
</td>
</tr>
</tbody></table>
<table style="width:98%;border:10px;border-spacing:10px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<table style="width:98%;border:10px;border-spacing:10px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2408.07050" id="MCG_journal">
<span class="papertitle">PSM: Learning Probabilistic Embeddings for Multi-scale Zero-Shot Soundscape Mapping</span>
</a>
<br>
<strong>Khanal Subash</strong>, Xing Eric, Sastry Srikumar, Dhakal Aayush, Xiong Zhexiao, Ahmad Adeel and Jacobs Nathan
<br>
<em>ACM Multimedia</em>, 2024
<br>
<a href="https://arxiv.org/pdf/2408.07050">arxiv</a> /
<a href="data/PSM.bib">bibtex</a> /
<a href="https://github.com/mvrl/PSM">code</a> /
<a href="PSM/index.html">project page</a>
<p></p>
<p>We develop a probabilistic, multi-scale, and metadata-aware embedding space that connects audio, text, and overhead imagery. This enables the creation of dynamic, multi-scale soundscape maps for any geographic region, along with uncertainty estimates for the mapping.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2309.10667.pdf" id="MCG_journal">
<span class="papertitle">Learning Tri-modal Embeddings for Zero-Shot Soundscape Mapping</span>
</a>
<br>
<strong>Khanal Subash</strong>, Sastry Srikumar, Dhakal Aayush and Jacobs Nathan
<br>
<em>BMVC</em>, 2023
<br>
<a href="https://arxiv.org/pdf/2309.10667.pdf">arxiv</a> /
<a href="data/GeoCLAP_supplementary.pdf">supplementary</a> /
<a href="data/geoclapBMVC2023.bib">bibtex</a> /
<a href="https://github.com/mvrl/geoclap">code</a>
<p></p>
<p>We learn a tri-modal embedding space between audio, text and overhead imagery. This enables us to create soundscape maps over any geographic region, using either audio or textual queries.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2411.00683" id="MCG_journal">
<span class="papertitle">TaxaBind: A Unified Embedding Space for Ecological Applications</span>
</a>
<br>
Sastry Srikumar, <strong>Khanal Subash</strong>, Dhakal Aayush, Ahmad Adeel and Jacobs Nathan
<br>
<em>WACV</em>, 2025
<br>
<a href="https://arxiv.org/pdf/2411.00683">arxiv</a> /
<a href="data/taxabind.bib">bibtex</a> /
<a href="https://github.com/mvrl/TaxaBind">code</a> /
<a href="https://vishu26.github.io/taxabind/index.html">project page</a>
<p></p>
<p>TaxaBind is a suite of multimodal models useful for downstream ecological tasks covering six modalities: ground-level image, geographic location, satellite image, text, audio, and environmental features.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2307.15904.pdf" id="MCG_journal">
<span class="papertitle">Sat2Cap: Mapping Fine-Grained Textual Descriptions from Satellite Images</span>
</a>
<br>
Dhakal Aayush, Ahmad Adeel, <strong>Khanal Subash</strong>, Sastry Srikumar, Kerner Hannah and Jacobs Nathan
<br>
<em>CVPRW (EarthVision), 2024, <strong>Best Paper Award</strong> </em>
<br>
<a href="https://arxiv.org/pdf/2307.15904.pdf">arxiv</a> /
<a href="data/sat2cap2023.bib">bibtex</a> /
<a href="https://github.com/mvrl/GeoClip/">code</a>
<p></p>
<p>We train a contrastive learning framework, Sat2Cap on a novel large scale dataset. This enables us to create maps using free-form textual descriptions.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/abs/2404.06637" id="MCG_journal">
<span class="papertitle">GeoSynth: Contextually-Aware High-Resolution Satellite Image Synthesis</span>
</a>
<br>
Sastry Srikumar, <strong>Khanal Subash</strong>, Dhakal Aayush, and Jacobs Nathan
<br>
<em>CVPRW (EarthVision)</em>, 2024
<br>
<a href="https://arxiv.org/abs/2404.06637">arxiv</a> /
<a href="data/geosynth.bib">bibtex</a> /
<a href="https://github.com/mvrl/GeoSynth">code</a>
<a href="https://vishu26.github.io/geosynth/index.html">project page</a>
<p></p>
<p>
This work presents GeoSynth, a diffusion-based model for synthesizing satellite images with global style and image-driven layout control.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2310.19168v1" id="MCG_journal">
<span class="papertitle">BirdSAT: Cross-View Contrastive Masked Autoencoders for Bird Species Classification and Mapping</span>
</a>
<br>
Sastry Srikumar, <strong>Khanal Subash</strong>, Di Huang, Dhakal Aayush and Jacobs Nathan
<br>
<em>WACV</em>, 2024
<br>
<a href="https://arxiv.org/pdf/2310.19168v1">arxiv</a> /
<a href="data/birdsat2024.bib">bibtex</a> /
<a href="https://github.com/mvrl/BirdSAT">code</a>
<p></p>
<p>
This work presents a flexible framework, with vector embedding and metric learning variants, that supports both species distribution mapping with fine-grained visual classification.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/abs/2404.11720" id="MCG_journal">
<span class="papertitle">GeoBind: Binding text, image, and audio through satellite images.</span>
</a>
<br>
Dhakal Aayush, <strong>Khanal Subash</strong>, Sastry Srikumar, Ahmad Adeel, Jacobs Nathan
<br>
<em>IGARSS </em>, 2024
<br>
<a href="https://arxiv.org/abs/2404.11720">arxiv</a> /
<a href="data/geobind.bib">bibtex</a>
<p></p>
<p>
This work presents a general framework that can be used to create an embedding space with any number of modalities by using satellite images as the binding element.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2312.08334.pdf" id="MCG_journal">
<span class="papertitle">LD-SDM: Language-Driven Hierarchical Species Distribution Modeling</span>
</a>
<br>
Sastry Srikumar, Xin Xing, Dhakal Aayush, <strong>Khanal Subash</strong>, Ahmad Adeel, and Jacobs Nathan
<br>
<em>preprint</em>, 2024
<br>
<a href="https://arxiv.org/pdf/2312.08334.pdf">arxiv</a> /
<a href="data/ldsdm.bib">bibtex</a>
<p></p>
<p>
We introduced a novel approach for species distribution modeling that uses a large-language model to generate a representation of species. This provides flexibility to generate range maps at different levels of the taxonomic hierarchy and for unseen species.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2407.09672" id="MCG_journal">
<span class="papertitle">Mixed-View Panorama Synthesis using Geospatially Guided Diffusion.</span>
</a>
<br>
Xiong Zhexiao, Xing Xin, Workman Scott, <strong>Khanal Subash</strong>, Jacobs Nathan
<br>
<em>preprint</em>, 2024
<br>
<a href="https://arxiv.org/pdf/2407.09672">arxiv</a> /
<a href="data/mixed.bib">bibtex</a>
<p></p>
<p>
This work introduces the task of mixed-view panorama synthesis, where the goal is to synthesize a novel panorama given a small set of input panoramas and a satellite image of the area.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://arxiv.org/pdf/2206.14841.pdf" id="MCG_journal">
<span class="papertitle">Causality for inherently explainable transformers: CAT-XPLAIN</span>
</a>
<br>
<strong>Khanal Subash</strong>, Brodie Benjamin, Xing Xin, Lin Ai-Ling and Jacobs Nathan
<br>
<em>CVPR Workshop</em>, 2022
<br>
<a href="https://arxiv.org/pdf/2206.14841.pdf">arxiv</a> /
<a href="data/catxplain2022.bib">bibtex</a> /
<a href="https://github.com/mvrl/CAT-XPLAIN">code</a>
<p></p>
<p>Add an extra special token (explainable token) into Vision Transformer (ViT), and train it to select the most important patches in the input image.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://ieeexplore.ieee.org/document/9761584" id="MCG_journal">
<span class="papertitle">Advit: Vision transformer on multi-modality pet images for alzheimer disease diagnosis</span>
</a>
<br>
Xing Xin, Liang Gongbo, Zhang Yu, <strong>Khanal Subash</strong>, Lin Ai-Ling and Jacobs Nathan.
<br>
<em>ISBI</em>, 2022
<br>
<a href="https://ieeexplore.ieee.org/document/9761584">paper</a> /
<a href="data/advit2022.bib">bibtex</a>
<p></p>
<p> Training ViT on 3D-to-2D converted multi-modal PET images achieves better Alzheimer's disease prediction.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://ieeexplore.ieee.org/document/9669730" id="MCG_journal">
<span class="papertitle">Alzheimer's Disease Classification Using Genetic Data</span>
</a>
<br>
<strong>Khanal Subash</strong>, Chen Jin, Jacobs Nathan and Lin Ai-Ling
<br>
<em>BIBM Workshop</em>, 2021
<br>
<a href="https://ieeexplore.ieee.org/document/9669730">paper</a> /
<a href="data/genetics2021.bib">bibtex</a> /
<a href="https://github.com/mvrl/ADNI_Genetics">code</a>
<p></p>
<p> Machine learning on different types of genetic data helps to identify candidate genes for Alzheimer's disease progression.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://ieeexplore.ieee.org/document/9554405" id="MCG_journal">
<span class="papertitle">Hierarchical Probabilistic Embeddings for Multi-View Image Classification</span>
</a>
<br>
Brodie Benjamin, <strong>Khanal Subash</strong>, Rafique Muhammad Usman, Greenwell Connor and Jacobs Nathan
<br>
<em>IGARSS</em>, 2021
<br>
<a href="https://ieeexplore.ieee.org/document/9554405">paper</a> /
<a href="data/probEmb2021.bib">bibtex</a>
<p></p>
<p> Learning a hierarchical, probabilistic embedding space allows one to achieve uncertainty estimate of feature distributions coming from sources with variable bands of information.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://ieeexplore.ieee.org/document/9383574" id="MCG_journal">
<span class="papertitle">Articulatory Comparison of L1 and L2 Speech for Mispronunciation Diagnosis</span>
</a>
<br>
<strong>Khanal Subash</strong>, Johnson Michael T. and Bozorg Narjess
<br>
<em>SLT</em>, 2021
<br>
<a href="https://ieeexplore.ieee.org/document/9383574">paper</a> /
<a href="data/slt2021.bib">bibtex</a>
<p></p>
<p> This paper compares the difference in articulatory patterns between native (L1) and non-native (L2) Mandarin speakers of English, for the purpose of providing an understanding of mispronunciation behaviors of L2 learners.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://ieeexplore.ieee.org/document/9587408" id="MCG_journal">
<span class="papertitle">Mispronunciation Detection and Diagnosis for Mandarin Accented English Speech</span>
</a>
<br>
<strong>Khanal Subash</strong>, Johnson Michael T., Soleymanpour Mohammad and Bozorg Narjes
<br>
<em>SpeD</em>, 2021
<br>
<a href="https://ieeexplore.ieee.org/document/9587408">paper</a> /
<a href="data/sped2021.bib">bibtex</a>
<p></p>
<p>Articulatory features improve the performance of Automatic Speech Recognition (ASR) based Mispronunciation Detection and Diagnosis (MDD) systems.</p>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<a href="https://uknowledge.uky.edu/cgi/viewcontent.cgi?article=1164&context=ece_etds" id="MCG_journal">
<span class="papertitle">Mispronunciation Detection and Diagnosis in Mandarin Accented English Speech</span>
</a>
<br>
<strong>Khanal Subash</strong>
<br>
<em>Theses and Dissertations--Electrical and Computer Engineering</em>, 2020
<br>
<a href="https://uknowledge.uky.edu/cgi/viewcontent.cgi?article=1164&context=ece_etds"> Thesis</a> /
<a href="data/thesis.bib">bibtex</a>
<p></p>
<p> The focus of this work was to analyse articulatory patterns of mispronunciation and design of ASR based MDD system.</p>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;"><tbody>
<tr>
<td style="padding:0px">
<br>
<p style="text-align:center;font-size:small;">
This website is modified from source code of <a href="https://github.com/jonbarron/jonbarron_website">John Barron's website</a>.
</p>
</td>
</tr>
</tbody></table>
</td>
</tr>
</table>
</body>
</html>