-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathresearch.html
201 lines (175 loc) · 8.73 KB
/
research.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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=900">
<title>Research</title>
<link rel="stylesheet" href="styles.css">
<link rel="icon" href="assets/icon.jpg">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Castoro"
rel="stylesheet">
<link href="https://fonts.googleapis.com/css?family=Noto+Sans:400,700,400italic,700italic"
rel="stylesheet">
<!-- <link href='http://fonts.googleapis.com/css?family=Lato:400,700,400italic,700italic' rel='stylesheet'> -->
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.15.3/css/all.min.css">
</head>
<body>
<table width="900" border="0" align="center" cellspacing="0" cellpadding="20">
<tr>
<td style="width:25%; vertical-align:middle; padding-right: 10px;">
<a href="index.html"><img src="assets/logo.png" height="100"></a>
</td>
<td style="width:15%; vertical-align:middle; text-align:center; padding-right: 5px;">
<a href="people.html" style="font-size: 22px; color:black">People</a>
</td>
<td style="width:15%; vertical-align:middle; text-align:center; padding-right: 10px;">
<a href="publications.html" style="font-size: 22px; color:black">Publications</a>
</td>
<td style="width:15%; vertical-align:middle; text-align:center; padding-right: 0px;">
<a href="research.html" style="font-size: 22px; color:black">Research</a>
</td>
<td style="width:15%; vertical-align:middle; text-align:center; padding-right: 0px;">
<a href="robots.html" style="font-size: 22px; color:black">Robots</a>
</td>
<td style="width:15%; vertical-align:middle; text-align:center; padding-right: 0px;">
<a href="join.html" style="font-size: 22px; color:black">Join</a>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Research Talks</heading>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<ul>
<li>2024 Oct: CMU RI Seminar Talk "Building Generalist Robots with Agility via Learning and Control: Humanoids and Beyond" (one hour) [<a href="https://youtu.be/Uym3Tr6t5TM?si=dJWQgO-mxbUlIwoQ" target="_blank"><i class="fas fa-globe"></i> recording</a>] [<a href="https://drive.google.com/file/d/18vldRta6Fp-tWPpft37OZO61EkED4Mh8/view?usp=sharing" target="_blank"><i class="far fa-file"></i> slides</a>]</li>
<li>2024 Sep: Georgia Tech IRIM Seminar Talk "Unifying Semantic and Physical Intelligence for Generalist Humanoid Robots" (one hour) [<a href="https://mediaspace.gatech.edu/media/1_tckra9zz" target="_blank"><i class="fas fa-globe"></i> recording</a>]</li>
<li>2023 Sep: New faculty lighting talk at CMU SCS (5 mins) [<a href="https://youtu.be/1MH-R6_UALw?si=isH5XlHQzg1719jy" target="_blank"><i class="fas fa-globe"></i> recording</a>]</li>
</ul>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<p>
This page is a bit outdated. Will update soon!
</p>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Safe Learning-based Nonlinear Control with Learned Robotic Agility</heading>
</td>
</tr>
<tr>
<td style="width:100%; vertical-align:middle">
<div class="image-container">
<img src='research/safe_learning_control.png' width="65%">
</div>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<p>Recent advances in machine learning beckon to applications in autonomous systems. However, for safety-critical settings, the learning system must interact with the rest of the autonomous system in a way that safeguards against catastrophic failures with guarantees. In addition, from computational and statistical standpoints, the learning system must incorporate prior knowledge for efficiency and generalizability. Leveraging control-theoretic tools and prior knowledge, we aim to develop learning-based control methods with both guarantees and new capabilities.
</p>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Offline Learning and Online Adaptation</heading>
</td>
</tr>
<tr>
<td style="width:100%; vertical-align:middle">
<div class="image-container">
<img src='research/adaptive.png' width="40%">
</div>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<p>Real-world robotic systems have to operate in unknown and dynamic environments where the decision-maker must fast adapt to uncertainties. For example, legged robot rescue and search necessitates traversing complicated terrain conditions. Deep learning has representation power but is often too slow to update onboard. On the other hand, adaptive control can update as fast as the feedback control loop with guarantees. Our goal is to develop offline and online algorithms that can effectively learn from offline data and efficiently fine-tune/adapt in real time.
</p>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Structured Reinforcement Learning and Control</heading>
</td>
</tr>
<tr>
<td style="width:100%; vertical-align:middle">
<div class="image-container">
<img src='research/structured.png' width="50%">
</div>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<p>Most RL algorithms (e.g., TRPO, SAC) are general for all tasks. In contrast, drastically different control methods are developed for different systems/tasks, and their successes highly rely on structures inside these systems/tasks. We seek to encode these structures and algorithmic principles into black-box RL algorithms, to make RL algorithms more data-efficient, robust, interpretable, and safe. We are particularly interested in hierarchical and superpositional RL and control approaches.
</p>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Learning and Control Theory: Towards a Unified Framework</heading>
</td>
</tr>
<tr>
<td style="width:100%; vertical-align:middle">
<div class="image-container">
<img src='research/theory.png' width="70%">
</div>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<p>There are many closely related concepts in machine learning and control communities, for instance, model-based RL and optimal control, online learning and adaptive control, domain randomization and robust control, online optimization and MPC, just to name a few. We seek to build interfaces and unified frameworks, which not only deepen fundamental connections between learning and control, but inspire new algorithms. One example of such connections is analyzing MPC's dynamic regret (a learning-theoretic metric).
</p>
</td>
</tr>
</table>
<br>
<table width="880" border="0" align="center" cellspacing="0" cellpadding="0">
<tr>
<td width="100%" valign="middle">
<heading>Swarm Intelligence</heading>
</td>
</tr>
<tr>
<td style="width:100%; vertical-align:middle">
<div class="image-container">
<img src='research/swarm.png' width="35%">
</div>
</td>
</tr>
<tr>
<td width="100%" valign="middle">
<p>Robot swarm learning and control present multiple new challenges, such as complex interactions between agents and dynamic topology. We aim to develop scalable and robust decision-making methods for multi-agent robotic systems, by leveraging properties like symmetry, locality, and invariance. We are also interested in how different types of robots interact (e.g., drone and legged robot), and human-robot teaming.
</p>
</td>
</tr>
</table>
<br>
<br>
<br>
</body>
</html>