Correlation-Weighted Multi-Reward Optimization for Compositional Generation
ECCV 2026
Sole first author
Text-to-Image
Reward Optimization
Graduate Student Researcher · Artificial Intelligence
Jungmyung Wi
I work on visual generative models, vision-language models, and domain transfer with a focus on reward optimization and alignment for compositional text-to-image generation.
5Papers
4Excellence prize awards
AIGenerative models & VLMs
About
Researcher in visual generation and alignment.
I am a master's student in Artificial Intelligence at Korea University and a researcher in UTL Lab. My research interests include image generative models, text-to-image diffusion models, vision-language models, and domain transfer. Looking ahead, I am interested in whether foundation models truly perceive and reason about the world, beyond simply improving them with supervised fine-tuning.
Research Interests
Text-to-image generation, diffusion models, VLMs, domain generalization, and unsupervised domain adaptation.
Future Work
World models, vision-language-action models, and reasoning in LLMs and mLLMs.
Education
Education & research experience
2024.03 — Present
Korea University
Master's Degree in Artificial Intelligence · Seoul, Republic of Korea
UTL Lab Researcher · Advisor: Prof. Donghyun Kim
2018.03 — 2024.02
Dankook University
Bachelor's Degree in Mobile System Engineering · Yongin, Republic of Korea
HAIL Lab Undergraduate Research Intern · Advisor: Prof. Dongjae Kim
Publications
Selected papers
Self-Improving Diffusion Classifiers with Minority Preference Optimization
ECCV 2026
Co-first author
Diffusion Classifier
GRPO
Delving into Pre-training for Domain Transfer: A Broad Study of Pre-training for Domain Generalization and Domain Adaptation
IJCV in 2026
Sole first author
Domain Transfer
Pre-training
Training-Free Label Space Alignment for Universal Domain Adaptation
EAAI in 2026
Co-author
UDA
Evaluating Vision Transformer for Deep Reinforcement Learning
KIISE 2023
Co-first author
ViT
Deep RL
Awards
Awards
2025.08
Samsung Collegiate Programming Challenge 2025 (AI)
Excellence Prize · Samsung Research
Developed an efficient and lightweight VLM for general VQA tasks.
2023.12
Bio Health Data AI Contest (Dentistry)
Excellence Prize · Hongik University, Samsung Medical Center, Kyung Hee University Dental Hospital
Classified oral health risk using CT image data in a closed environment.
2022.12
Space Radio Disaster Prediction AI Competition
Excellence Prize · National Radio Research Agency
Built classification and regression models using space radio signals.
2022.07
Dacon Seoul Bike Rental Prediction
Excellence Prize · Yongin City in Republic of Korea
Predicted bike rental demand using feature engineering and machine learning models.
Skills
Skills & coursework
Programming
Python
PyTorch
C
Research Topics
Computer Vision
Text-to-Image
Diffusion Models
VLM
Domain Transfer
Reinforcement Learning
Coursework
Machine Learning
Deep Learning
Linear Algebra
Probability & Statistics
Languages
Korean · Native
English · Intermediate
Contact
Contact
Feel free to reach out for research discussions, collaboration, or project questions.