Youngtaek Oh

Contact: {firstname}.{lastname} [at] kaist.ac.kr

yto2.jpg

I am a Ph.D. student advised by Prof. In So Kweon and Prof. Junmo Kim at KAIST, and was a research intern at LG AI Reseach, working with Dr. Pyunghwan Ahn and Dr. Jinhyung Kim.

My research centers on developing data-efficient and robust recognition systems that address limitations and biases in image and video data. I am currently focusing on enhancing vision-language models to improve compositional reasoning and multi-modal understanding for more robust and comprehensive recognition systems.

Publications

  1. fsc-clip-teaser.png
    Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality
    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2024
    Oral presentation
  2. vl_compo.png
    Exploring the Spectrum of Visio-Linguistic Compositionality and Recognition
    Youngtaek Oh, Pyunghwan Ahn, Jinhyung Kim, Gwangmo Song, Soonyoung LeeIn So KweonJunmo Kim
    CVPR Workshop on ‘What is Next in Multimodal Foundation Models?’ (MMFM), 2024
  3. 2023_cviu_badapters.png
    Empirical study on using Adapters for debiased Visual Question Answering
    Computer Vision and Image Understanding (CVIU), 2023
  4. nice_teaser.png
    NICE 2023 Zero-shot Image Captioning Challenge
    Taehoon Kim, Pyunghwan Ahn, Sangyun Kim, Sihaeng Lee, Mark Marsden, Alessandra Sala, Seung Hwan Kim, Honglak Lee, Kyounghoon Bae, and 33 more authors
    preprint, 2023
  5. self_sufficient.png
    Self-Sufficient Framework for Continuous Sign Language Recognition
    International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2023
    Oral presentation, Top 3% recognition of all accepted papers
  6. signing_outside.png
    Signing Outside the Studio: Benchmarking Background Robustness for Continuous Sign Language Recognition
    British Machine Vision Conference (BMVC), 2022
  7. daso.png
    DASO: Distribution-Aware Semantics-Oriented Pseudo-Label for Imbalanced Semi-Supervised Learning
    Youngtaek OhDong-Jin KimIn So Kweon
    IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022
  8. ksl_guide.png
    KSL-Guide: A Large-scale Korean Sign Language Dataset Including Interrogative Sentences for Guiding the Deaf and Hard-of-Hearing
    Soomin Ham, Kibaek ParkYoungjoon JangYoungtaek Oh, Seokmin Yun, Sukwon Yoon, Chang Jo KimHan-Mu ParkIn So Kweon
    International Conference on Automatic Face and Gesture Recognition (FG), 2021
  9. sideguide.png
    SideGuide: A Large-scale Sidewalk Dataset for Guiding Impaired People
    {Kibaek ParkYoungtaek Oh, Soomin HamKyungdon Joo}*, Hyokyoung Kim, Hyoyoung KumIn So Kweon (*: equally contributed)
    International Conference on Intelligent Robots and Systems (IROS), 2020

Honors and Awards

Jun 2023 Top 3% Recognition Certificates, ICASSP 2023   [Certificate]
May 2023 2nd place in NICE Challenge ($5000), NICE Workshop at CVPR 2023   [Press] [Code]
Oct 2022 Outstanding reviewer award, ECCV 2022.