-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
394 lines (335 loc) · 15.5 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
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="One-Shot Imitation under Mismatched Execution">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>One-Shot Imitation under Mismatched Execution</title>
<!-- Google tag (gtag.js) -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-QR88D4MJ0H"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'G-QR88D4MJ0H');
</script>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/icon.jpg">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">One-Shot Imitation under Mismatched Execution</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://kushal2000.github.io">Kushal Kedia*</a>,</span>
<span class="author-block">
<a href="https://portfolio-pdan101.vercel.app/">Prithwish Dan</a>,</span>
<!-- <span class="author-block">
<a href="https://portal.cs.cornell.edu/people/">Angela Chao</a>,</span>
<span class="author-block">
<a href="https://portal.cs.cornell.edu/people/">Maximus Adrian Pace</a>,</span> -->
<span class="author-block">
<a href="https://www.sanjibanchoudhury.com/">Sanjiban Choudhury</a>
</div>
<!-- write Cornell Univerity in red next line -->
<br>
<!-- <hr style="height:0px; visibility:hidden;" /> -->
<img src="./static/images/cornell-university-logo.png" width="40%"></img>
<!-- <div class="is-size-3 publication-authors">
<span class="author-block">Cornell University</span>
</div> -->
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2409.06615"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<span class="link-block">
<a href="https://arxiv.org/abs/2409.06615"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
<!-- twitter Link. -->
<span class="link-block">
<a href="TODO"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-x-twitter"></i>
</span>
<span>Summary TODO</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- reduce line spacing below -->
<!-- Display a big video next -->
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-3" >
Retrieval based algorithm that trains policies to imitate demonstrator actions under one-shot initiation
</h1>
<br>
<div class="publication-video">
<!-- add a link to youtube video using iframe -->
<p >Video TODO</p>
<!-- </video> -->
</div>
</div>
</div>
</div>
</div>
</section>
<!-- <section class="hero is-light is-small">
<div class="hero-body" has-text-centered>
<div class="container" >
<div class="column has-text-centered">
<h2 class="title is-3">MOSAIC enables flexibility, recoverability, and safety</h2>
<br>
<div id="results-carousel" class="carousel results-carousel">
<div class="item item-steve">
<video poster="" id="steve" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/recoverability.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-chair-tp">
<video poster="" id="chair-tp" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/safety.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-shiba">
<video poster="" id="shiba" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/flexibility.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-steve">
<video poster="" id="steve" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/recoverability.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-chair-tp">
<video poster="" id="chair-tp" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/safety.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-shiba">
<video poster="" id="shiba" autoplay controls muted loop playsinline height="100%">
<source src="./static/videos/flexibility.mp4"
type="video/mp4">
</video>
</div>
</div>
<br>
</div>
</div>
</div>
</section> -->
<section class="section">
<div class="container is-max-desktop">
<!-- Dataset -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Abstract</h2>
<!-- <div class="publication-video"> -->
<p style="text-align: left;">
Human demonstrations as prompts are a powerful way to program robots to do long-horizon manipulation tasks.
However, translating these demonstrations into robot-executable actions presents significant challenges due to execution mismatches in movement styles and physical capabilities. Existing methods either depend on robot-demonstrator paired data, which is infeasible to scale, or rely too heavily on frame-level visual similarities that often break down in practice. To address these challenges, we propose RHyME, a novel framework that automatically aligns robot and demonstrator task executions using optimal transport costs. Given long-horizon robot demonstrations, RHyME synthesizes semantically equivalent demonstrator videos by retrieving and composing short-horizon demonstrator clips. This approach facilitates effective policy training without the need for paired data. We demonstrate that RHyME outperforms a range of baselines across cross-embodiment datasets, showing a 52% increase in task recall over prior cross-embodiment learning methods.
</p>
<!-- </div> -->
</div>
</div>
<br>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Training Overview</h2>
<!-- <div class="publication-video"> -->
<!-- align left -->
<!-- <p>image TODO</p> -->
<img src="./resources/fig_intro_new.png" style="width: 90%">
<br>
<br>
<!-- write three bullet points next using <ul> -->
<li style="text-align: left;"> <b>Visual Encoder</b>: Maps robot and demonstrator videos to a shared embedding space </li>
<li style="text-align: left;"> <b>RHyME</b>: Retrieve the nearest human snippet in embedding space to each robot video segment, then learn a policy based on combined human_video+robot_action and robot_video+robot_action pairs </li>
<!-- </div> -->
</div>
</div>
<br>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Visual Encoder</h2>
<!-- <div class="publication-video"> -->
<!-- align left -->
<img src="./resources/fig_tsne.jpeg">
<br>
<br>
<p style="text-align: left;">
The visual encoder aligns human demonstration encodings and robot demonstration encodings visually, temporally, and at a task level
</p>
<li style="text-align: left;"> <b>Visual Alignment</b>: Cluster images based on image features using SwAV </li>
<li style="text-align: left;"> <b>Temporal Alignment</b>: Drive images closer together in time together in embedding space and vice versa via Time Contrastive Loss. </li>
<li style="text-align: left;"> <b>Task Alignment</b>: WHEN some paired robot-demonstrator data is available, align human and robot videos together in embedding space. </li>
<!-- </div> -->
</div>
</div>
<br>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">RHyME</h2>
<!-- <div class="publication-video"> -->
<!-- align left -->
<!-- <img src="./static/images/hmf.png"> -->
<!-- <img src="./static/images/HMF_video.gif"> -->
<img src="./static/images/algo_img.png" style="width: 60%">
<br>
<br>
<p style="text-align: left;">
Given a collected robot video with actions, we imagine a corresponding human video performing exactly the same task. This is done by retrieving the nearest snippet of human video in dataset in embedding space to each snippet of robot video. This creates a set of paired human-robot-trajectory triplet data.With the set of paired human-robot-trajectory triplet data, a policy is trained to predict action based on (state, human_demo_embedding, action) pairs and (state, robot_demo_embedding, action) pairs.
</p>
<!-- </div> -->
</div>
</div>
<br>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Evaluations</h2>
<!-- <div class="publication-video"> -->
<!-- align left -->
<!-- <img src="./static/images/vs_approach_pic.png"> -->
<li style="text-align: left;"><b>Training Data</b>:
<br>
<img src="./static/images/train_data_1_place_holder.png" style="width: 45%; display: block; margin: 0 auto">
<br>
<img src="./static/images/train_data_2_place_holder.png" style="width: 45%; display: block; margin: 0 auto">
<!-- Light -> Pot -> Cloth, Pot -> Drawer -> Cloth -->
</li>
<li style="text-align: left;"><b>Test Data</b>:
<img src="./static/images/test_data_place_holder.png" style="width: 45%; display: block; margin: 0 auto">
</li>
<p> TODO: Lots of Videos </p>
<!-- Create an interactive section below with three
buttons corresponding to playing of one of three gifs -->
<!-- </div> -->
</div>
</div>
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Evaluation Results</h2>
<!-- <div class="publication-video"> -->
<!-- align left -->
<!-- <img src="./static/images/hmf.png"> -->
<br>
<br>
<p style="text-align: left;">
TODO put in the table
</p>
<!-- </div> -->
</div>
</div>
<!-- <br> -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column" id="Paper">
<h2 class="title is-3">Paper TODO</h2>
<div class="paper-thumbnail">
<a href="https://arxiv.org/abs/2402.18796.pdf">
<img class="layered-paper-big" width="100%" src="./resources/preview_mosaic.png" alt="Paper thumbnail"/>
</a>
</div>
<div class="paper-info">
<!-- <a href="https://openreview.net/forum?id=rxlokRzNWRq"><h3>ManiCast: Collaborative Manipulation with Cost-Aware Human Forecasting</h3></a>
<p>Kushal Kedia, Prithwish Dan, Atiksh Bhardwaj, Sanjiban Choudhury</p> -->
<h3 class="title is-4" style="text-align: center;">BibTex</h3>
<pre style="overflow-x:hidden; text-wrap:wrap; white-space: pre-wrap;"><code>@misc{wang2024mosaic,
title={MOSAIC: A Modular System for Assistive and Interactive Cooking},
author={Huaxiaoyue Wang and Kushal Kedia and Juntao Ren and Rahma Abdullah and Atiksh Bhardwaj and Angela Chao and Kelly Y Chen and Nathaniel Chin and Prithwish Dan and Xinyi Fan and Gonzalo Gonzalez-Pumariega and Aditya Kompella and Maximus Adrian Pace and Yash Sharma and Xiangwan Sun and Neha Sunkara and Sanjiban Choudhury},
year={2024},
eprint={2402.18796},
archivePrefix={arXiv},
primaryClass={cs.RO}
}
</code></pre>
</div>
</div>
<!-- <div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@inproceedings{
kedia2023manicast,
title={ManiCast: Collaborative Manipulation with Cost-Aware Human Forecasting},
author={Kushal Kedia and Prithwish Dan and Atiksh Bhardwaj and Sanjiban Choudhury},
booktitle={7th Annual Conference on Robot Learning},
year={2023},
url={https://openreview.net/forum?id=rxlokRzNWRq}
}</code></pre>
</div> -->
</div>
</section>
<br>
<footer class="footer">
<div class="container">
<!-- <div class="content has-text-centered">
<a class="icon-link"
href="./static/videos/nerfies_paper.pdf">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/keunhong" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div> -->
<div class="columns is-centered">
<div class="column is-8">
<div class="content has-text-centered">
<p>
Website template borrowed from <a
href="https://github.com/nerfies/nerfies.github.io">NeRFies</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>