-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathindex.html
executable file
·440 lines (371 loc) · 17.1 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
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
<!DOCTYPE HTML>
<html lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>Zhi Wang</title>
<meta name="author" content="Zhi Wang">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="shortcut icon" href="images/nju_icon.png" 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;">
Zhi Wang
</p>
<p>I'm an Associate Professor at <a href="https://www.nju.edu.cn/">Nanjing University</a> in Nanjing, China. I received the Ph.D. degree from City University of Hong Kong and the bachelor's degree from Nanjing University. I was a visiting scholar at University of New South Wales and Institute of Automation, Chinese Academy of Sciences.
</p>
<p style="text-align:center">
<a href="mailto:[email protected]">Email</a> /
<a href="https://scholar.google.com/citations?user=cRXlxYcAAAAJ&hl=en">Google Scholar</a> /
<a href="https://github.com/NJU-RL/">Github</a> /
<a href="publication.html">Publication</a>
</p>
</td>
<td style="padding:2.5%;width:25%;max-width:25%">
<a href="images/ZhiWang.jpg"><img style="width:100%;max-width:100%;object-fit: cover; border-radius: 50%;" alt="profile photo" src="images/ZhiWang.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:16px;width:100%;vertical-align:middle">
<h2>Research</h2>
<p>
I'm interested in reinfocement learning algorithms and applications.
Specifically, I work on how learning algorithms can scale RL agents to i) dynamic environments, ii) offline settings, and iii) multi-agent systems, allowing them to autonomously adapt to i) non-stationary task distributions, ii) non-interactive scenarios, and iii) cooperative or competitive task assignments, facilitating RL's deployment in real-world domains.
</p>
<p>
Recently, I work on leveraging foundation models in decision-making problems, exploring ideas of language agents, in-context RL, and embodied intelligence.
</p>
</td>
</tr>
</tbody></table>
<table style="width:100%; margin:0 auto; border:0; border-spacing:0; padding:16px;">
<table style="width:100%; margin:0 auto; border:0; border-spacing:0; padding:16px;"><tbody>
<tr>
<td>
<h3>Generalization in RL</h3>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tbody>
<tr onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/Meta-DT.jpg' width=100%>
</div>
<img src='images/Meta-DT.jpg' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://arxiv.org/abs/2410.11448">
<span class="papertitle">Meta-DT: Offline Meta-RL as Conditional Sequence Modeling with World Model Disentanglement</span>
</a>
<br>
<strong>Zhi Wang</strong>, Li Zhang, Wenhao Wu, Yuanheng Zhu, Dongbin Zhao, Chunlin Chen
<br>
<em>NeurIPS</em>, 2024
<br>
<a href="https://github.com/NJU-RL/Meta-DT">code</a>
/
<a href="paper/2024-NeurIPS-Meta-DT.pdf">paper</a>
<p></p>
<p>
We leverage the sequential modeling ability of the transformer architecture and robust task representation learning via world model disentanglement to achieve efficient generalization in offline meta-RL.
</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 onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/llirl.jpg' width=100%>
</div>
<img src='images/llirl.jpg' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://ieeexplore.ieee.org/abstract/document/9353402">
<span class="papertitle">Lifelong Incremental Reinforcement Learning with Online Bayesian Inference</span>
</a>
<br>
<strong>Zhi Wang</strong>, Chunlin Chen, Daoyi Dong
<br>
<em>IEEE Transactions on Neural Networks and Learning Systems</em>, 2022
<br>
<a href="https://github.com/HeyuanMingong/llirl">code</a>
/
<a href="paper/LLIRL.pdf">paper</a>
<p></p>
<p>
We develop a lifelong RL agent that can incrementally adapt its behaviors to dynamic environments, via maintaining an ever-expanding policy library with online Bayesian inference.
</p>
</td>
</tr>
</tbody>
</table>
<table style="width:100%; margin:0 auto; border:0; border-spacing:0; padding:16px;"><tbody>
<tr>
<td>
<h3>Multi-Agent RL</h3>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tbody>
<tr onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/acorm.jpg' width=100%>
</div>
<img src='images/acorm.jpg' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://arxiv.org/abs/2312.04819">
<span class="papertitle">Attention-Guided Contrastive Role Representations for Multi-Agent Reinforcement Learning</span>
</a>
<br>
Zican Hu, Zongzhang Zhang, Huaxiong Li, Chunlin Chen, Hongyu Ding, <strong>Zhi Wang</strong>*
<br>
<em>ICLR</em>, 2024
<br>
<a href="https://github.com/NJU-RL/ACORM">code</a>
/
<a href="paper/2024-ICLR-ACORM.pdf">paper</a>
<p></p>
<p>
Our main insight is to learn a compact role representation that can capture complex behavior patterns of agents, and use that role representation to promote behavior heterogeneity, knowledge transfer, and skillful coordination across agents.
</p>
</td>
</tr>
<tr onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/mixrts.png' width=100%>
</div>
<img src='images/mixrts.png' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://arxiv.org/abs/2209.07225">
<span class="papertitle">MIXRTs: Toward Interpretable Multi-Agent Reinforcement Learning Via Mixing Recurrent Soft Decision Trees</span>
</a>
<br>
Zichuan Liu, Yuanyang Zhu, <strong>Zhi Wang</strong>*, Yang Gao, Chunlin Chen
<br>
<em>IEEE Transactions on Pattern Analysis and Machine Intelligence</em>, 2025
<br>
<a href="">code</a>
/
<a href="">paper</a>
<p></p>
<p>
We propose a novel architecture based on differentiable soft decision trees to tackle the tension between model interpretability and learning performance in MARL domains, paving the way for interpretable and high-performing MARL systems.
</p>
</td>
</tr>
</tbody>
</table>
<table style="width:100%; margin:0 auto; border:0; border-spacing:0; padding:16px;"><tbody>
<tr>
<td>
<h3>RL Applications and Others</h3>
</td>
</tr>
</tbody></table>
<table style="width:100%;border:0px;border-spacing:0px;border-collapse:separate;margin-right:auto;margin-left:auto;">
<tbody>
<tr onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/osp.jpg' width=100%>
</div>
<img src='images/osp.jpg' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://ieeexplore.ieee.org/document/8668561">
<span class="papertitle">Reinforcement Learning-Based Optimal Sensor Placement for Spatiotemporal Modeling</span>
</a>
<br>
<strong>Zhi Wang</strong>, Han-Xiong Li, Chunlin Chen
<br>
<em>IEEE Transactions on Cybernetics</em>, 2020
<br>
<a href="paper/RL_OSP.pdf">paper</a>
<p></p>
<p>
For the first time, we introduce an RL-based method to tackle the optimal sensor placement problem for modeling distributed parameter systems.
</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 onmouseout="ever_stop()" onmouseover="ever_start()">
<td style="padding:16px;width:20%;vertical-align:middle">
<div class="one">
<div class="two" id='ever_image'>
<img src='images/iwfit.jpg' width=100%>
</div>
<img src='images/iwfit.jpg' width=100%>
</div>
<script type="text/javascript">
function ever_start() {
document.getElementById('ever_image').style.opacity = "1";
}
function ever_stop() {
document.getElementById('ever_image').style.opacity = "0";
}
ever_stop()
</script>
</td>
<td style="padding:8px;width:80%;vertical-align:middle">
<a href="https://ojs.aaai.org/index.php/AAAI/article/view/4709">
<span class="papertitle">Better Fine-Tuning via Instance Weighting for Text Classification</span>
</a>
<br>
<strong>Zhi Wang</strong>, Wei Bi, Yan Wang, Xiaojiang Liu
<br>
<em>AAAI</em>, 2019
<br>
<a href="paper/IW_Fit.pdf">paper</a>
/
<a href="paper/IW_Fit_supp.pdf">supp</a>
<p></p>
<p>
we propose an Instance Weighting based Fine-tuning (IW-Fit) method, which revises the fine-tuning stage to improve the classification accuracy on the target domain when a pre-trained model from the source domain is given.
</p>
</td>
</tr>
</tbody>
</table>
<table style="width:100%; margin:0 auto; border:0; border-spacing:0; padding:16px;"><tbody>
<tr>
<td>
<h2>Miscellanea</h2>
</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 align="center" style="padding:16px;width:20%;vertical-align:middle">
<div class="colored-box" style="background-color: #edd892;">
<h2>Teaching</h2>
</div>
</td>
<td style="padding:8px;width:80%;vertical-align:center">
<<a href="teaching.html">Deep Reinforcement Learning</a>>, for postgraduates
<br>
<Digital Circuits>, for undergraduates
</td>
</tr>
<tr>
<td align="center" style="padding:16px;width:20%;vertical-align:middle">
<div class="colored-box" style="background-color: #c6b89e;">
<h2>Academic Service</h2>
</div>
</td>
<td style="padding:8px;width:80%;vertical-align:center">
Associate Editor, Special Sessions, IEEE SMC 2023/2022/2021, IEEE ICNSC 2020
<br>
Reviewer: ICML/NeurIPS/ICLR/CVPR/AAAI/ECAI, IEEE TPAMI/TNNLS/TCYB/TSYS/TMECH/JAS
</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 align="center" style="padding:16px;width:20%;vertical-align:middle">
<hr>
<footer>
<p>Last updated: <script type="text/javascript">document.write(document.lastModified);</script>.   
<!-- Default Statcounter code for My personal homepage
http://heyuanmingong.io -->
<script type="text/javascript">
var sc_project=12838348;
var sc_invisible=0;
var sc_security="fdf6cfe3";
var scJsHost = "https://";
document.write("<sc"+"ript type='text/javascript' src='" +
scJsHost+
"statcounter.com/counter/counter.js'></"+"script>");
</script>
<noscript><div class="statcounter"><a title="free hit
counter" href="https://statcounter.com/"
target="_blank"><img class="statcounter"
src="https://c.statcounter.com/12838348/0/fdf6cfe3/0/"
alt="free hit counter"
referrerPolicy="no-referrer-when-downgrade"></a></div></noscript>
<!-- End of Statcounter Code -->
</p>
</footer>
</td>
</tr>
</tbody></table>
</td>
</tr>
</table>
</body>
</html>