M.S. in Computer Science
(Sogang University)
π Hiroshima, Japan
βοΈ E-Mail jeongskim512@gmail.com
π Google Scholar Link
Hello! I am Independent Researcher. I recently obtained my M.S. in Computer Science from Sogang University, where I focused on computer vision and deep learning.
My research primarily focuses on building highly efficient vision models and diagnosing/resolving structural or algorithmic bottlenecks to bridge the gap between theoretical models and practical hardware deployment.
And also, I am interested in Representation Editing and Parameter-Efficient Fine-Tuning to learn or unlearn specific knowledge from model, and currently research how to isolate and eliminate domain bias, achieving domain-agnostic semantic alignment.
Driven by curiosity and a passion for community-driven AI, I am always open to new research collaborations, discussions, and academic opportunities. Feel free to reach out if you would like to discuss potential projects or simply connect!
π Looking for Opportunities: I am currently seeking a Ph.D. program to join in Fall 2027.
CUST : Clustered Unit-level Similarity Transformer for Lightweight Image Super-Resolution
Jeongsoo Kim
European Conference on Computer Vision (ECCV), 2026.
[Paper] [Code]
LMLT : Low-to-high Multi-Level Vision Transformer for Lightweight Image Super-Resolution
Jeongsoo Kim , Jongho Nang, Junsuk Choe*
Workshop of International Conference on Computer Vision (ICCV), 2025.
[Paper] [Code]