Youngwook Kim

I am a Postdoc researcher at Department of Electrical and Computer Engineering, Seoul National University, with advisor Jungwoo Lee.

I received my Ph.D degree from Department of Eletrical and Computer Engineering, Seoul National University, in 2024.

I received my bachelor's degree from Department of Civil and Environmental Engineering, Seoul National University, in 2018. During this period, I received a two-year National Science & Technology Scholarship.

Email / Google Scholar / Github / Linkedin / Resume

Research

I'm interested in computer vision, learning from noisy labels, multi-label learning, weakly supervised learning, and remote sensing.

Large Loss Matters in Weakly Supervised Multi-Label Classification
Youngwook Kim*, Jae Myung Kim*, Zeynep Akata, Jungwoo Lee
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022
[GitHub] / [arXiv] / [Poster] / [Slide]

By utilizing memorization effect, we can efficiently train multi-label classification model with small number of partially annotated labels.

Bridging the Gap between Model Explanations in Partially Annotated Multi-label Classification
Youngwook Kim, Jae Myung Kim, Jieun Jeong, Cordelia Schmid, Zeynep Akata, Jungwoo Lee
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023
[GitHub] / [arXiv] / [Project page] / [Poster] / [Slide] / [YouTube]

By investigating the difference in model explanation, we devise a simple but effective solution to compensate the damage by false negative labels.

Misc

Invited Talk	University of Seoul, June 2022 Kakao enterprise, July 2022 AIIS Spring retreat (Oral presenation), May 2023 [Presentation video] (in Korean) Smart city seminar, Seoul National University, Department of Civil and Environmental Engineering, September 2023
TA Experience	Introduction to deep learning, Fall 2018 Information theory, Spring 2020
Reviewer Service	Conferences : CVPR, ICCV, ECCV Journals : IEEE Transactions on Multimedia

The format of this website is borrowed from Jon Barron.