Hyounguk Shon

PhD student at KAIST SIIT Lab.

prof_pic.jpg

Affiliation: KAIST

Daejeon, South Korea

Currently a PhD student at KAIST SIIT Lab. My supervisor is Prof. Junmo Kim and I am also advised by Prof. Yunho Jeon. I am interested in machine learning, deep learning, and computer vision. My recent works involve continual learning and unlearning, noisy label learning, multi-modal contrastive learning, and scalable hyperparameter search. Previously, I was a research intern at LG AI Research mentored by Dr. Janghyeon Lee.

news

Jul 22, 2023 In our new paper for ICCV 2023, we explore machine unlearning in the context of transfer learning. We introduce a transfer learning strategy titled Disposible Transfer Learning (DTL). DTL filters out extra information during fine-tuning, which is useful for limiting the risk of exposing the pre-trained model when publishing the expert model. Our goal is to help continue open-source AI in the era of foundation models via striking a balance between transparency and security of ownership.

selected publications

  1. WACV
    SFLD: Reducing the content bias for AI-generated Image Detection
    Seoyeon Gye*, Junwon Ko*, Hyounguk Shon*, Minchan Kwon, and Junmo Kim
    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (Oral presentation), 2025
    HTML
  2. IJCAI Workshop
    Neural Vicinal Risk Minimization: Noise-robust Distillation for Noisy Labels
    Hyounguk Shon, Seunghee Koh, Yunho Jeon, and Junmo Kim
    In Proceedings of the IJCAI-2024 AISafety Workshop, 2024
    HTML
  3. AAAI
    FRED: Towards a Full Rotation-Equivariance in Aerial Image Object Detection
    Chanho Lee, Jinsu Son, Hyounguk Shon, Yunho Jeon, and Junmo Kim
    In Proceedings of the AAAI Conference on Artificial Intelligence, 2024
    HTML Poster
  4. ICCV
    Disposable Transfer Learning for Selective Source Task Unlearning
    Seunghee Koh, Hyounguk Shon, Janghyeon Lee, Hyeong Gwon Hong, and Junmo Kim
    In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), Oct 2023
    HTML
  5. ECCV
    DLCFT: Deep Linear Continual Fine-Tuning for General Incremental Learning
    Hyounguk Shon, Janghyeon Lee, Seung Hwan Kim, and Junmo Kim
    In European Conference on Computer Vision, Oct 2022
    HTML Poster
  6. CVPR
    Joint Negative and Positive Learning for Noisy Labels
    Youngdong Kim, Juseung Yun, Hyounguk Shon, and Junmo Kim
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun 2021
    HTML
  7. NeurIPS
    UniCLIP: Unified Framework for Contrastive Language-Image Pre-training
    Janghyeon Lee*, Jongsuk Kim*, Hyounguk Shon, Bumsoo Kim, Seung Hwan Kim, Honglak Lee, and Junmo Kim
    In Advances in Neural Information Processing Systems, Jun 2022
    HTML
  8. ICRA
    Lightweight Monocular Depth Estimation via Token-Sharing Transformer
    Dong-Jae Lee*, Jae Young Lee*, Hyunguk Shon, Eojindl Yi, Yeong-Hun Park, Sung-Sik Cho, and Junmo Kim
    In IEEE International Conference on Robotics and Automation, Jun 2023
    HTML
  9. ECCV
    On the Angular Update and Hyperparameter Tuning of a Scale-Invariant Network
    Juseung Yun, Janghyeon Lee, Hyounguk Shon, Eojindl Yi, Seung Hwan Kim, and Junmo Kim
    In European Conference on Computer Vision, Jun 2022
    HTML