Suchae Jeong
I am an undergraduate student majoring in Computer Science and Engineering at Dongguk University.
My research focuses on aligning robot foundation models using reinforcement learning, with particular interests in:
Vision-Language-Action Model, World Model, and Multi-Agent System
Reinforcement Learning from Human (or AI) Feedback
AI Alignment and Safety
Currently, I conduct research at the RISE Lab at KAIST, where I explore robot intelligence with ongoing work in multi-view spatial reasoning and human-robot interaction.
With a passion for learning, I embrace challenges as opportunities for growth and innovation.
Publications
(C: conference / J: journal / D: domestic / P: preprint / *: equal contribution / ^: equal advising)
2025
[C2] Culture-TRIP: Culturally-Aware Text-to-Image Generation with Iterative Prompt Refinement [pdf]
Suchae Jeong*, Inseong Choi*, Youngsik Yun, Jihie Kim
Annual Conference of the Nations of the Americas Chapter of the ACL (NAACL). 2025
2024
[D1] Easy-Read Conversion using LLM: Investigating Sentence Transformation Ability in Terms of Readability and Semantic Similarity [pdf]
Suchae Jeong, Jaehyeong Hwang, Hyesu Kim, Minji Lee, Yeji Lee, Jaehun Lee
Annual Conference of Korea Information Processing Society (ACK). 2024
[C1] Improving LLM Classification of Logical Errors by Integrating Error Relationship into Prompts [pdf]
Yanggyu Lee, Suchae Jeong, Jihie Kim
International Conference on Intelligent Tutoring Systems (ITS). 2024 (Best Paper Award)
Experience
RISE Lab [May. 2025 ~ Present]
Student Researcher, Korea Advanced Institute of Science and Technology (KAIST), Seoul, South Korea
May. 2025 ~ Present
Research Area:
Robot intelligence, including multi-view spatial reasoning, while expanding into human-robot interaction.
Advisor: Kimin Lee
Config Intelligence [Feb. 2025 ~ Present]
Data Intern, Seoul, South Korea
Main Role:
Tailored existing robot control interfaces (Polymetis, libfranka, and ROS 2) to enhance usability for project-specific teleoperation workflows
Developed and evaluated metrics to assess and iteratively improve inverse kinematics and inverse dynamics control
Built data preprocessing and annotation pipelines to process and label teleoperation data for model training
Implemented a scheduling system for automated model inference and integrated it with a dedicated model server
Supervisor: Sungdong Kim
Co-advisors: Kimin Lee, Minjoon Seo, and Hyungmok Son
Machine Learning Lab [Aug. 2023 ~ May. 2025]
Student Researcher, Dongguk University, Seoul, South Korea
Research Area:
Fairness in AI, Cultural Alignment of Generative Models [C2]
LLM Application [C1] [D1]
Advisor: Jihie Kim
Education
B.S. [Mar. 2020 ~ Present]
Department of Computer Science and Engineering
Mar. 2020 ~ Present
GPA: 4.41/4.5 (3rd place out of 77)
Dongguk University, Seoul, South Korea
Advisor: Kangwoo Lee