Sehwan Park
M.S. Student MIIL, Korea UniversityI am an M.S. student at the Multimodal Interactive Intelligence Laboratory (MIIL), Korea University, advised by Prof. Paul Hongsuck Seo. I also work closely with Prof. Seung Wook Kim.
I work on generative AI for vision, from diffusion distillation to controllable and compositional generation. My long-term goal is to build generative models that learn from limited data by capturing the physical and relational structure of the visual world, so that they can serve as world models for machines that perceive, imagine, and act.
News
- [Aug. 2026] Compositional SVG Generation is accepted to EMNLP 2026.
- [Nov. 2025] Random Conditioning is selected as a Qualcomm Innovation Fellowship Korea Finalist.
- [May 2025] Cross-Modal Watermarking is accepted to Interspeech 2025.
- [Feb. 2025] Random Conditioning is accepted to CVPR 2025.
Publications
-
CVPR
CVPR 2025 (Qualcomm Innovation Fellowship Korea Finalist) -
Interspeech
EMNLP
Education
M.S. in Computer Science and Engineering, Korea University
2025.03 – 2027.08 (expected)
Multimodal Interactive Intelligence Lab (MIIL) · Advisor: Prof. Paul Hongsuck Seo
B.S. in Intelligence Mechatronics Engineering, Sejong University
2019.03 – 2025.02
MINES Lab · Advisor: Prof. Hyung Seok Kim · Graduated Summa Cum Laude (GPA 4.31/4.5)
Awards & Honors
Qualcomm Innovation Fellowship Korea — Finalist
Nov. 2025
For Random Conditioning (CVPR 2025).
2nd-rank Honor Scholarship (GPA 4.5/4.5)
Spring 2024
2nd-rank Honor Scholarship (GPA 4.5/4.5)
Spring 2023
1st Place, Grand Prize — Regional Connection SW Presentation, Seoul Campus Town
Sep. 2020
1st-rank Honor Scholarship (GPA 4.5/4.5)
Fall 2019
Teaching
Online Lecturer, Practice: Diffusion Models
Sep. 2025
Recorded for Samsung Electronics
Teaching Assistant, Samsung Electronics AI Expert Program
Aug. 2025
Action Recognition and Video Understanding · Visual Question Answering
Powered by Jekyll and Minimal Light theme.