Younghun Kim의 프로필 사진

Younghun Kim

Master in GGGS at KAIST

I am currently pursuing a Master's degree in GGGS at KAIST, advised by Prof.Changick Kim.

While theoretical knowledge is critical, its real-world application is just as essential. I am deeply passionate about exploring the diverse applications of deep learning across various fields. In particular, I am keenly interested in ensuring the robustness and stability of deep learning models, especially when applied in complex, real-world scenarios. Through my work, I aim to bridge the gap between groundbreaking theory and practical implementation, ensuring that deep learning innovations are both powerful and reliable.

My work explores the following, but not limited to:

  • AI Safety AI-Safety
    • Deepfake Detection
    • Image Forensic
    • Bias
  • Language Models Language
  • Medical AI Medical AI

Education

  • M.S. in GGGS, KAIST, 2026 (Expected)
  • B.S. in CS, HGU, 2024
    (Major GPA: 4.33/4.5)
  • GVCS, 2018
    (Student Council President)

Contact

younghun1664@kaist.ac.kr

dudgjs4248@gmail.com

291, Daehak-ro, Yuseong-gu, Daejeon, Republic of Korea 34141

Publications

2025

SELFI
Deepfake Detection
SELFI: Selective Fusion of Identity for Generalizable Deepfake Detection
Younghun Kim, Minsuk Jang, Myung-Joon Kwon, Wonjun Lee, and Changick Kim
[P2] arXiv preprint arXiv:2506.17592 2025
ReFT
Language Models (Safe Tuning)
Refusal-Feature-guided Teacher for Safe Finetuning via Data Filtering and Alignment Distillation
Seokil Ham, Yubin Choi, Seungju Cho, Yujin Yang, Younghun Kim, and Changick Kim
[P1] arXiv preprint arXiv:2506.07356 2025
MBS
Deepfake Detection
A Learning Framework for Mitigating the Problem of Deepfake Detection Models Over-focusing on Specific Generation Techniques
Younghun Kim, and Changick Kim
[D5] The IEIE Summer Conference 2025
LiTE
Liteweight
LiTE ViT: Leaky Integrate-and-Fire Token Efficient Vision Transformer
Minsuk Jang, Younghun Kim, Byeongjun Park, and Changick Kim
[D4] The IEIE Summer Conference 2025

2024

FRIDAY
Deepfake Detection
FRIDAY: Mitigating Unintentional Facial Identity in Deepfake Detectors Guided by Facial Recognizers
Younghun Kim, Myung-Joon Kwon, Wonjun Lee, and Changick Kim
[C1] IEEE International Conference on Visual Communication and Image Processing (VCIP Oral Presentation) 2024
PECI
Medical Segmentation
PECI-Net: Bolus segmentation from video fluoroscopic swallowing study images using preprocessing ensemble and cascaded inference
Dougho Park, Younghun Kim, Harim Kang, Junmyeoung Lee, Jinyoung Choi, Taeyeon Kim, Sangeok Lee, Seokill Son, Minsol Kim, and Injung Kim
[J1] Computers in Biology and Medicine 2024 (IF: 7.7)
Diffusion Model Patching
Deepfake Detection
Improving the Generalization performance of Deepfake Detection through a Partial Image Manipulation Augmentation Technique
Younghun Kim, Myung-Joon Kwon, Wonjun Lee, and Changick Kim
[D1] The IEIE Summer Conference 2024
FRIDAY
IFL
Ensembling Pretrained Models for Enhanced Generalization in Synthetic Image Detection
Myung-Joon Kwon, Wonjun Lee, Younghun Kim, and Changick Kim
[D2] The IEIE Summer Conference 2024
FRIDAY
Bias
Combatting Bias: Corrective Measures through Adversarial Training in Neural Networks
Wonjun Lee, Myung-Joon Kwon, Younghun Kim, and Changick Kim
[D3] The IEIE Summer Conference 2024

Projects

  • Bolus Segmentation in VFSS (2022 - 2023)
    Pohang Stroke and Spine Hospital
  • Hate Speech AI Bot Detection (2025)
    Cyber Security Research Center

Presentations

  • 2025 KAIST Green Growth and Sustainability Workshop
    Oral Presentation at Climate Policy, Trust, Energy Transitions
  • 2024 IEEE International Conference on Visual Communication and Image Processing
    Oral Presentation at Security/Privacy

Awards

  • 2023-1 Capstone Festival [Grand Prize]
  • 2023 Gyeongbuk Hackaton 4th [Encouragement prize]
  • 2022, 2023 Academic Excellent Scholarship
  • 2022 SW Festival - Smart Application [Encouragement prize]
  • 2021 SW Festival - Smart Application [Excellence prize]

Experiences

  • (2024 - Present) Computational Intelligence Lab
  • (2022 - 2023) Deep Learning Lab: VFSS Medical Image Segmentation
  • (2023) C Programming TA
  • (2023) Operating System Course TA
  • (2021 - 2022) WALab: Web and App Developer