shashacks.github.io/index.html at master · shashacks/shashacks.github.io · GitHub

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
<!DOCTYPE html>
<html>
    <head>
        <meta charset="UTF-8">
        <meta name="author" content="Ju-Seung Byun">
        <meta name="description" content="Ju-Seung Byun is a Postdoctoral Researcher at LG AI Researcher">
        <meta name="keywords" content="Ju-Seung Byun,Reinforcment Learning,Machine Learning,Inha,USC,OSU">
        <title>Ju-Seung Byun</title>
        <link rel="shortcut icon" type="image/x-icon" href="./assets/images/ohio.png"/>
        <link rel="stylesheet" href="./assets/css/style.css">
        <link rel="stylesheet" href="./assets/css/font.css">
        <link rel="stylesheet" href="./assets/academicons-1.9.2/css/academicons.min.css"/>
        <script src="https://kit.fontawesome.com/0a608ee6b2.js" crossorigin="anonymous"></script>
        <meta name="viewport" content="width=device-width; initial-scale=1.0;">
    </head>
    <body>
    <header>
        <div class="header_container">
            <div class="header_title">
                <a href="https://shashacks.github.io/">Ju-Seung Byun</a>
            </div>
            <div class="header_subtitle">
                <!-- <a href="./assets/pdfs/Curriculum_Vitae_2024.pdf">CV</a> -->
                <!-- <a href="/">Miscellaneous</a> -->
                <!-- <a href="/">Publications</a> -->
                <!-- <a href="https://shashacks.github.io/">Home</a> -->
            </div>
        </div>
    </header>

    <div class="outer_container">
        <div class="profile">
            <img src="./assets/images/profile_image.jpg" class="profile_image" alt="Ju-Seung Byun">
            <div class="profile_name">Ju-Seung Byun</div>
            <div class="profile_div">Postdoctoral Researcher</div>
            <div class="profile_div">LG AI Research</div>
            <!-- <div class="profile_div"></div> -->
            <div class="profile_div">Advanced Agent Lab</div>
            <div class="profile_incon_div">
                <a href="mailto:byun.83@osu.edu">
                    <i class="fa-solid fa-envelope"></i></i>
                </a>
                <a href="https://scholar.google.com/citations?user=yKcK_BMAAAAJ">
                    <i class="ai ai-google-scholar-square"></i>
                </a>
                <a href="https://github.com/shashacks">
                    <i class="fa-brands fa-github"></i></i>
                </a>
                <a href="https://linkedin.com/in/ju-seung-byun-1a76b01b9">
                    <i class="fa-brands fa-linkedin"></i></i>
                </a>
            </div>
            <ul class="profile_incon_ul">
                <li >
                    <a>
                        <i class="fa-solid fa-location-pin profile_icon"></i>
                    &nbsp;Ann Arbor, Michigan
                    </a>
                </li>
                <li>
                    <a href="mailto:byun.83@osu.edu">
                        <i class="fa-solid fa-envelope"></i></i>
                        &nbsp;Email
                    </a>
                </li>
                <li>
                    <a href="https://scholar.google.com/citations?user=yKcK_BMAAAAJ">
                        <i class="ai ai-google-scholar-square"></i>
                    &nbsp;Google Scholar
                    </a>

                </li>
                <li>
                    <a href="https://github.com/shashacks">
                        <i class="fa-brands fa-github"></i></i>
                    &nbsp;Github
                    </a>
                </li>
                <li>
                    <a href="https://linkedin.com/in/ju-seung-byun-1a76b01b9">
                        <i class="fa-brands fa-linkedin"></i></i>
                    &nbsp;Linkedin
                    </a>
                </li>
            </ul>

        </div>
        <div class="main_container">
            <h1>About me</h1>
            <p>
                My name is Ju-Seung Byun. I am a Postdoctoral Researcher at LG AI Research, working with Prof. <a href="https://scholar.google.com/citations?user=fmSHtE8AAAAJ&hl=ko">Honglak Lee</a>.
                I received my Ph.D. in Computer Science and Engineering from The Ohio State University, advised by Prof. <a href="https://aperrault.github.io/">Andrew Perrault</a>.
                My research interests are in deep reinforcement learning, with a focus on GUI agents.
            </p>
            <p>
                Before my Ph.D., I earned my Master’s degree from the University of Southern California (2019) and my Bachelor’s degree from Inha University (2017) in South Korea, both in Computer Science.
            </p>
            <br><br>
            <h1>Publications</h1>


            <div class="publication">
                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/srl.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>Symmetric Reinforcement Learning Loss for Robust Learning on Diverse Tasks and Model Scales </b></span><br>
                        <span class="author"><b>Ju-Seung Byun</b>, Andrew Perrault  </span><br>
                        <span class="description"><b>ICML 2025</b> / <a href="https://arxiv.org/pdf/2405.17618">Paper</a> / <a href="https://github.com/shashacks/Symmetric_RL">Code</a></span>
                    </div>
                </div>

                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/Normality-Guided Distributional Reinforcement Learning for Continuous Control.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>Normality-Guided Distributional Reinforcement Learning for Continuous Control</b></span><br>
                        <span class="author"><b>Ju-Seung Byun</b>, Andrew Perrault </span><br>
                        <span class="description"><b>TMLR 2025</b> / <a href="https://arxiv.org/abs/2208.13125">Paper</a> / <a href="https://github.com/shashacks/MC_CLT_Sumbission">Code</a></span>
                    </div>
                </div>

                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/wg2.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>Fine-Tuning Text-to-Speech Diffusion Models Using Reinforcement Learning with Human Feedback</b></span><br>
                        <span class="author">Jingyi Chen, <b>Ju-Seung Byun</b>, Micha Elsner, Andrew Perrault </span><br>
                        <span class="description"><b>Interspeech 2025</b> / <a href="https://arxiv.org/abs/2405.14632">Paper</a> </span><br>
                    </div>
                </div>

                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/ares.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>ARES: Alternating Reinforcement Learning and Supervised Fine-Tuning for Enhanced Multi-Modal Chain-of-Thought Reasoning Through Diverse AI Feedback </b></span><br>
                        <span class="author"><b>Ju-Seung Byun*</b>, Jiyun Chun*, Jihyng Kil, Andrew Perrault  </span><br>
                        <span class="description"><b>EMNLP 2024</b> / <a href="https://arxiv.org/abs/2407.00087">Paper</a> / <a href="https://github.com/Amyyyyeah/ARES">Code</a></span>
                    </div>
                </div>

                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/salsabot.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>SalsaBot: Towards a Robust and Generalizable Embodied Agent </b></span><br>
                        <span class="author">Chan Hee Song*, Jiaman Wu* <b>Ju-Seung Byun</b>, Zexin Xu, Vardaan Pahuja, Goonmeet Bajaj, Samuel Stevens, Ziru Chen, Yu Su </span><br>
                        <span class="description"><b>Embodied AI Workshop at CVPR 2023</b> / <a href="https://embodied-ai.org/papers/2023/10.pdf">Paper (Short)</a> </span><br>
                        <span class="description"><b>Alexa Prize SimBot Challenge Proceedings 2023</b> / <a href="https://assets.amazon.science/ea/b7/0895f4e0468680903efcfd67e795/salsabot-report-1.pdf">Paper (Long)</a> </span>
                    </div>
                </div>


                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/Training Transition Policies via Distribution Matching for Complex Tasks.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>Training Transition Policies via Distribution Matching for Complex Tasks</b></span><br>
                        <span class="author"><b>Ju-Seung Byun</b>, Andrew Perrault </span><br>
                        <span class="description"><b>ICLR 2022</b> / <a href="https://openreview.net/forum?id=6vkzF28Hur8">Paper</a> / <a href="https://github.com/shashacks/IRL_Transition">Code</a> /  <a href="https://www.youtube.com/watch?v=xXBYhxk-mcg">Video</a> </span>
                    </div>
                </div>

                <div class="publication_container">
                    <div class="publication_img">
                        <img src="./assets/images/Proximal Policy Gradient: PPO with Policy Gradient.png">
                    </div>
                    <div class="publication_text">
                        <span class="title"><b>Proximal Policy Gradient: PPO with Policy Gradient</b></span><br>
                        <span class="author"><b>Ju-Seung Byun</b>, Byungmoon Kim, Huamin Wang </span><br>
                        <span class="description"><b>arXiv 2020</b> / <a href="https://arxiv.org/abs/2010.09933">Paper</a> / <a href="https://github.com/shashacks/PPG_Research">Code</a></span>
                    </div>
                </div>
            </div>
            <!-- <h1>Miscellaneous</h1>
            <ul>
                <li>hi</li>
                <li>hi2</li>
            </ul> -->
        </div>
    </div>
    <footer>
        <div>
            This website is motivated from <a href="https://youngwoon.github.io/">Youngwonn Lee</a> and <a href="https://jiajunwu.com/"> Jiajun Wu</a>
        </div>
    </footer>
    </body>
</html>