Minsoo Kang
I am an M.S. candidate in the Department of Artificial Intelligence at Korea University, advised by Professor Seong-Whan Lee, and a student researcher in the Center for Artificial Intelligence at KIST co-advised by Senior Researcher Suhyun Kim . I am especially interested in robust model training, such as data augmentation (Combining and Generative model-based) and data-free continual learning. My previous research had focused on a real-world applicable method by utilizing reinforcement learning.
[github] [cv] [LinkedIn]
Contact: minsoo [dot] kang0918 [at] gmail [dot] com
Research Interest
Data Augmentation, Continual Learning, Model Inversion (Diffusion-based), Regularization and Computer Vision
Education
Korea University, Seoul, Korea
M.S. in Department of Artificial Intelligence (4.44 / 4.5)
2022.03. - 2024.08. (Expected)
Soongsil University, Seoul, Korea
B.S. in Electronic Engineering (4.30 / 4.5)
2016.03 - 2022.02. (Summa Cum Laude, 1/97)
Publications
International
Catch-up Mix: Catch-up Class for Struggling Filters in CNN
Minsoo Kang*, Minkoo Kang*, Suhyun Kim (*: Equal contribution)
Association for the Advancement of Artificial Intelligence (AAAI), 2024. [arxiv] [blog] [code]
GuidedMixup: An Efficient Mixup Strategy Guided by Saliency Maps
Minsoo Kang, Suhyun Kim
Association for the Advancement of Artificial Intelligence (AAAI, Oral), 2023. [arxiv] [blog] [code]
Mixup Mask Adaptation: Bridging the gap between input saliency and representations via attention mechanism in feature mixup
Minsoo Kang, Minkoo Kang, Seong-Whan Lee, Suhyun Kim
Image and Vision Computing (SCIE, IF: 4.7), 2023. [paper]
Domestic
Search Algorithm for Genetic Variants that Increases the Effectiveness of Patient Treatment using Reinforcement Learning and Proportional Hazards Model
Minkoo Kang, Minsoo Kang, Dongjin Kim, Suhyun Kim
The Journal of Korean Institute of Next Generation Computing (KINGPC, KCI), 2022. [code]
Local Route Selection and Speed Control of Autonomous Vehicles Based on Decentralized Deep Reinforcement Learning
Minsoo Kang, Ilkyu Yi, Yunsik Jo, Ohsoon Shin
ICT Express Summer Conference (KICS), 2021. (Best Paper Award) [code]
Deep Reinforcement Learning Based Velocity Control for Autonomous Vehicle in Circular Road
Minsoo Kang, Hyunju Lim, Gihong Lee, Minhae Kwon
ICT Express Fall Conference (KICS), 2020. [code]
Deep Reinforcement Learning Based Autonomous Vehicles for Traffic Flow Control
Gihong Lee, Hyunju Lim, Minsoo Kang, Minhae Kwon
ICT Express Fall Conference (KICS), 2020. (Best Paper Award) [code]
Experience
Student Researcher, at Center for Artifical Intelligence, KIST
- Data augmentation, Continual learning, Debiasing Representation and Model Inversion
2022.03 - Present (Adviser: Dr. Suhyun Kim)
Research Intern, at Center for Artifical Intelligence, KIST
- Data augmentation and Knowledge Distillation
2021.03 - 2021.12. (Adviser: Dr. Suhyun Kim)
Research Intern, at Department of Smart Data, AI Research Center, ETRI
- Decentralized deep reinforcement learning based traffic signal control
- Traffic flow control with real-world applicable constraints
2021.01 - 2021.02. (Adviser: Dr. Munyung Jung) [code]
Research Intern, at BMIL, Soongsil University
- Deep reinforcement learning based autonomous driving
- Data based velocity and flow control
2020.07 - 2020.09. (Adviser: Prof. Minhae Kwon) [code]
More details are in my CV.
