-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
223 lines (198 loc) · 11.3 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
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<meta name="description" content="Hi, this is Huayu Chen, a student from Tsinghua University. 嗨，我叫陈华玉，目前是清华大学计算机系的博士生。">
<meta name="keywords" content="Huayu Chen, ChenHuayu, chenhuayu, 陈华玉, chenDRAG, reinforcement learning, generative models, Tsinghua, 清华, department of computer science, 计算机">
<link rel="stylesheet" href="./jemdoc.css" type="text/css">
<title>Huayu Chen's Homepage</title>
<link rel="shortcut icon" href="favicon.ico" type="image/x-icon">
<link rel="icon" href="favicon.ico" type="image/x-icon">
<meta name="baidu-site-verification" content="kZzZTRvzUQ" />
</head>
<body>
<div id="layout-content" style="margin-top:25px">
<table><tbody><tr>
<td width="670">
<div id="toptitle"><h1>Huayu Chen (陈华玉) </h1></div>
<h3>PhD student</h3>
<p>
Room 1-509, FIT Building <br>
Dept. of Computer Science and Technology <br>
Tsinghua University <br>
Beijing, China, 100084. <br>
Email: chenhuay17[AT]gmail[DOT]com <br>
<!-- <a href="https://github.com/ChenDRAG">[GitHub]</a> <br> -->
<a href="https://scholar.google.com.au/citations?user=0FBCHc4AAAAJ&hl=en&oi=ao">[Google Scholar]</a> <br>
</p>
</td>
<td><img src="./images/me.jpg" border="0" width="200"></td>
</tr></tbody></table>
<h2>Biography</h2>
<p> I am a fourth-year PhD student of <a href="http://ml.cs.tsinghua.edu.cn/index.html">TSAIL Group</a> in the <a href="https://www.cs.tsinghua.edu.cn/csen/">Department of Computer Science and Technology</a>, <a href="https://www.tsinghua.edu.cn">Tsinghua University</a>, advised by <a href="http://ml.cs.tsinghua.edu.cn/~jun/">Prof. Jun Zhu</a> and <a href="https://www.suhangss.me/">Prof. Hang Su</a>.
Currently, I am also a research intern at Nvidia <a href="https://research.nvidia.com/labs/dir/"> Deep Imagination Research group </a> in the San Francisco Bay Area.
Previously, I received my B.S. degree from the Department of Automation of Tsinghua University in July 2021.
I spent a wonderful time at the Digital Media Lab at Tsinghua University, advised by <a href="http://www.liuyebin.com/">Prof. Yebin Liu</a> in the field of AIGC from Oct 2018 to May 2019.
I have also been a research intern at <a href="https://fuxi.163.com/">Netease's Fuxi AI Lab </a> and <a href="https://ailab.bytedance.com/sdk">ByteDance's AI Lab</a> respectively in 2021 and 2020.
</p>
<p>
Currently, my research interests lie primarily in the area of <strong>deep reinforcement learning</strong> and <strong>deep generative models</strong>.
My lifelong goal is to build a scalable, impenetrable, and adaptable decision-making engine that could relieve human from tedious tasks and elevate their work efficiency.
My current progress includes authoring <strong>Tianshou</strong>: A highly modularized deep reinforcement learning library <a href="https://github.com/thu-ml/tianshou"><img alt="GitHub stars" src="https://img.shields.io/github/stars/thu-ml/tianshou?style=social" class="github"></a>, designing large-scale Online RL system for mastering MOBA games (see <a href="#competitions"> Competitions</a>), and bridging the gap between RL theories and generative modeling methods such as LLM/diffusion.
</p>
<h2>Selected Publications</h2>
<strong> RL Infra: </strong>
<ul>
<li>
<a href='https://arxiv.org/abs/2107.14171'> Tianshou: A Highly Modularized Deep Reinforcement Learning Library </a> <br>
<!-- <font color="#FF0000">Oral (Accept rate~1.7%)</font> <br> -->
Jiayi Weng*, <strong>Huayu Chen*</strong>, Dong Yan, Kaichao You, Alexis Duburcq, Minghao Zhang, Yi Su, Hang Su, Jun Zhu <br>
Journal of Machine Learning Research <strong>(JMLR)</strong> <br>
<a href='https://github.com/thu-ml/tianshou'>[Code: <font color="#FF0000">8.2k Stars</font>]</a>
</li>
</ul>
<strong> RL for LLM: </strong>
<ul>
<li>
<a href='http://arxiv.org/abs/2402.05369'>Noise Contrastive Alignment of Language Models with Explicit Rewards </a> <br>
<strong>Huayu Chen</strong>, Guande He, Lifan Yuan, Ganqu Cui, Hang Su, Jun Zhu <br>
Annual Conference on Neural Information Processing Systems <strong>(NeurIPS 2024)</strong> <br>
<a href='https://github.com/thu-ml/Noise-Contrastive-Alignment'>[code]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2503.15558'>Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning </a> <br>
<strong>Nvidia Group</strong> (Contributing to VLM RL training) <br>
<a href='https://research.nvidia.com/labs/dir/cosmos-reason1/'>[project page]</a>
<a href='https://github.com/nvidia-cosmos/cosmos-reason1'>[code]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2502.01456'>Process Reinforcement through Implicit Rewards </a> <br>
Ganqu Cui, Lifan Yuan, Zefan Wang, Hanbin Wang, Wendi Li, Bingxiang He, Yuchen Fan, Tianyu Yu, Qixin Xu, Weize Chen, Jiarui Yuan, <strong>Huayu Chen</strong>, Kaiyan Zhang, Xingtai Lv, Shuo Wang, Yuan Yao, Xu Han, Hao Peng, Yu Cheng, Zhiyuan Liu, Maosong Sun, Bowen Zhou, Ning Ding <br>
<a href='https://arxiv.org/abs/2502.01456'>[Preprint]</a>
<a href='https://github.com/PRIME-RL/PRIME'>[Code: <font color="#FF0000">1.4k Stars</font>]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2412.01981'>Free Process Rewards without Process Labels </a> <br>
Lifan Yuan, Wendi Li, <strong>Huayu Chen</strong>, Ganqu Cui, Ning Ding, Kaiyan Zhang, Bowen Zhou, Zhiyuan Liu, Hao Peng <br>
<a href='https://arxiv.org/abs/2412.01981'>[Preprint]</a>
<a href='https://github.com/PRIME-RL/ImplicitPRM'>[code]</a>
</li>
</ul>
<strong> RL for Vision (Diffusion & AR): </strong>
<ul>
<li>
<a href='http://arxiv.org/abs/2501.15420'>Visual Generation Without Guidance </a> <br>
<strong>Huayu Chen*</strong>, Kai Jiang*, Kaiwen Zheng, Jianfei Chen, Hang Su, Jun Zhu <br>
<a href='http://arxiv.org/abs/2501.15420'>[Preprint]</a>
<a href='https://github.com/thu-ml/GFT'>[code]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2410.09347'>Toward Guidance-Free AR Visual Generation via Condition Contrastive Alignment </a> <br>
<strong>Huayu Chen</strong>, Hang Su, Peize Sun, Jun Zhu <br>
International Conference on Learning Representations <strong>(ICLR 2025)</strong> <br>
<font color="#FF0000">Oral (Accept rate~1.8%)</font> <br>
<a href='https://github.com/thu-ml/CCA'>[code]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2304.12824'>Contrastive Energy Prediction for Exact Energy-Guided Diffusion Sampling in Offline Reinforcement Learning </a> <br>
Cheng Lu*, <strong>Huayu Chen*</strong>, Jianfei Chen, Hang Su, Chongxuan Li, Jun Zhu <br>
International Conference on Machine Learning <strong>(ICML 2023)</strong> <br>
<a href='https://github.com/ChenDRAG/CEP-energy-guided-diffusion'>[code]</a>
<a href='./images/icml2023_poster_CEP_fix2_20230715172121.pdf'>[poster]</a>
</li>
</ul>
<strong> RL for Embodied AI (Diffusion Policy): </strong>
<ul>
<li>
<a href='https://arxiv.org/abs/2410.07864'>RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation </a> <br>
Songming Liu, Lingxuan Wu, Bangguo Li, Hengkai Tan, <strong>Huayu Chen</strong>, Zhengyi Wang, Ke Xu, Hang Su, Jun Zhu <br>
International Conference on Learning Representations <strong>(ICLR 2025)</strong> <br>
<a href='https://rdt-robotics.github.io/rdt-robotics/'>[project page]</a>
<a href='https://github.com/thu-ml/RoboticsDiffusionTransformer'>[code]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2407.09024'>Aligning Diffusion Behaviors with Q-functions for Efficient Continuous Control </a> <br>
<strong>Huayu Chen</strong>, Kaiwen Zheng, Hang Su, Jun Zhu <br>
Annual Conference on Neural Information Processing Systems <strong>(NeurIPS 2024)</strong> <br>
<!-- <a href='https://github.com/thu-ml/Noise-Contrastive-Alignment'>[code]</a> -->
</li>
<li>
<a href='https://arxiv.org/abs/2310.07297'>Score Regularized Policy Optimization through Diffusion Behavior </a> <br>
<strong>Huayu Chen</strong>, Cheng Lu, Zhengyi Wang, Hang Su, Jun Zhu <br>
International Conference on Learning Representations <strong>(ICLR 2024)</strong> <br>
<a href='https://github.com/thu-ml/SRPO'>[code]</a>
<a href='./images/iclr2024-poster-v1mine.pdf'>[poster]</a>
</li>
<li>
<a href='https://arxiv.org/abs/2209.14548'>Offline Reinforcement Learning via High-Fidelity Generative Behavior Modeling</a> <br>
<strong>Huayu Chen</strong>, Cheng Lu, Chengyang Ying, Hang Su, Jun Zhu <br>
International Conference on Learning Representations <strong>(ICLR 2023)</strong> <br>
<a href='https://github.com/ChenDRAG/SfBC'>[code]</a>
<a href='./images/iclr2023-poster-v1mine.pdf'>[poster]</a>
</li>
</ul>
* indicates co-first authors.
<h2 id="competitions">Competitions</h2>
<ul>
<li>
<strong>First place (two years in a row) in Tencent's multi-agent RL competition of Honor of Kings (王者荣耀), <font color="#FF0000">final win rate: 99.2%</font> </strong>, 2021-2023 <br>
<a href='https://www.tencent.com/en-us/articles/2201392.html'>[news]</a>
<a href='https://aiarena.tencent.com/aiarena/zh/match/aiarena-competition-3rd'>[webpage]</a>
</li>
<li>
Second place in DJI's Robomaster Sim2Real Challenge, ICRA 2022
</li>
<li>
First place in the 30th International Design Contest (IDC Robocon 2019, MIT), 2019
</li>
<li>
First place in the 20th Electronic Design Competition at Tsinghua University, 2018
</li>
<li>
First place in the 1st Artificial Intelligence Challenge in Tsinghua University, 2017
</li>
</ul>
<h2>Honors & Awards</h2>
<ul>
<li>
HUAWEI-Tsinghua Scholarship, 2023
</li>
<li>
'84' Future Innovation Scholarship, 2023
</li>
<li>
<strong>Outstanding Undergraduate in Beijing, 2021</strong>
</li>
<li>
BaoGang Scholarship (Awarded to ~500 students in China every year), 2021
</li>
<li>
Student Of The Year, in Dept. of Automation, Tsinghua University, 2020
</li>
<li>
<strong>China National Scholarship, 2019</strong>
</li>
<li>
Excellence Award for Technological Innovation, Tsinghua University, 2019
</li>
<li>
<strong>'129' Scholarship (Highest honor for 2nd year students in the Dept. of Automation, Tsinghua University), 2018</strong>
</li>
<li>
1st Prize in the 35th China Regional College Students Physics Competition, 2018
</li>
<li>
1st Prize in the 30th National Physics Olympiad, 2016
</li>
</ul>
<h2>Services</h2>
<li>Reviewer for ICLR, NeurIPS, ICML, AISTATS, AAAI, etc.</li>
<li>President of Student Association of Science and Technology, Dept. of Automation, Tsinghua University, 2020-2021</li>
<h2>Teaching</h2>
2023 Spring, TA in <strong>Statistical Learning Theory and Applications</strong>, instructed by <a href="http://ml.cs.tsinghua.edu.cn/~jun/">Prof. Jun Zhu</a><br>
</div>
<div id="footer">
<div id="footer-text"></div>
</div>
© 2024 Huayu Chen
</body>
</html>