Jaehyuk Jang
Ph.D. Student
Computational Intelligence Lab
KAIST
jhyuk(at)kaist.ac.kr, 1120jhyuk(at)gmail.com
Education
  • KAIST
    KAIST
    Mar. 2022 - Present
    Ph.D. in Electrical Engineering
    Daejeon, Korea
    Advised by Prof. Changick Kim
  • KAIST
    KAIST
    Mar. 2020 - Feb. 2022
    M.S. in Electrical Engineering
    Daejeon, Korea
    Advised by Prof. Jong-Hwan Kim
  • Hanyang University
    Hanyang University
    Mar. 2016 - Feb. 2020
    B.S. in Electrical Engineering
    Seoul, Korea
    • About Me

    I'm Jaehyuk Jang, a Ph.D. student at Korea Advanced Institute of Science and Technology (KAIST), advised by Prof. Changick Kim.

    News
    2026
    Aug
    Four paper accepted to EMNLP 2026 (1 Main, 3 Findings).
    Jun
    One paper accepted to Interspeech 2026.
    May
    One paper accepted to ICML 2026 as a Spotlight.
    Apr
    One paper accepted to ACL 2026 (1 Findings).
    Jan
    One paper accepted to ICASSP 2026.
    2025
    May
    One paper accepted to ACL 2025 Findings.
    Jan
    One paper published in Pattern Recognition.
    2024
    Mar
    One paper published in Pattern Recognition Letters.
    2023
    Dec
    One paper accepted to ICASSP 2024.
    Sep
    One paper published at VCIP 2023.
    2021
    Jun
    One paper published at IROS 2021.
    Publications
    * Equal contribution, † Corresponding author
    2026
    Who Wins the Conflict? Mechanistic Interpretability of Text Bias in Audio LLMs

    H. Cho, S. Yoo, Jaehyuk Jang, Changick Kim, J. S. Chung

    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026

    [arXiv]

    Who Wins the Conflict? Mechanistic Interpretability of Text Bias in Audio LLMs

    H. Cho, S. Yoo, Jaehyuk Jang, Changick Kim, J. S. Chung

    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026

    [arXiv]

    SubT: Subspace Tuning for Few-shot Generalization of Audio-Language Models

    Jaehyuk Jang, K. Ko, W. Lee, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv] [Project page]

    SubT: Subspace Tuning for Few-shot Generalization of Audio-Language Models

    Jaehyuk Jang, K. Ko, W. Lee, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv] [Project page]

    Where Identity Lives: Localized, Retain-Free Identity Unlearning in Multimodal Large Language Models

    K. Ko*, Jaehyuk Jang*, W. Lee*, H.-S. Kim, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv]

    Where Identity Lives: Localized, Retain-Free Identity Unlearning in Multimodal Large Language Models

    K. Ko*, Jaehyuk Jang*, W. Lee*, H.-S. Kim, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv]

    AIM: Anchor Identity Features, Then Match for Multimodal Large Language Model Unlearning

    W. Lee*, Jaehyuk Jang*, K. Ko*, H.-S. Kim, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv]

    AIM: Anchor Identity Features, Then Match for Multimodal Large Language Model Unlearning

    W. Lee*, Jaehyuk Jang*, K. Ko*, H.-S. Kim, Changick Kim

    Findings of Conference on Empirical Methods in Natural Language Processing (EMNLP Findings), 2026

    [arXiv]

    Acoustic Prompting via Stage-wise Modulation for Few-Shot Learning in Audio Language Models

    H. Cho, Jaehyuk Jang, Changick Kim†, J. S. Chung†

    Interspeech, 2026

    [arXiv] [Project page]

    Acoustic Prompting via Stage-wise Modulation for Few-Shot Learning in Audio Language Models

    H. Cho, Jaehyuk Jang, Changick Kim†, J. S. Chung†

    Interspeech, 2026

    [arXiv] [Project page]

    Multimodal Self-Attention Network with Temporal Alignment for Audio-Visual Emotion Recognition

    I. Koo, Y. Seong, M. Son, Jaehyuk Jang, Changick Kim

    IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026

    [Paper] [arXiv]

    Multimodal Self-Attention Network with Temporal Alignment for Audio-Visual Emotion Recognition

    I. Koo, Y. Seong, M. Son, Jaehyuk Jang, Changick Kim

    IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026

    [Paper] [arXiv]

    Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models

    S. Ham, Jaehyuk Jang, W. Lee, Changick Kim

    Proceedings of the International Conference on Machine Learning (ICML), 2026 Spotlight

    [Paper] [arXiv]

    Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models

    S. Ham, Jaehyuk Jang, W. Lee, Changick Kim

    Proceedings of the International Conference on Machine Learning (ICML), 2026 Spotlight

    [Paper] [arXiv]

    Generalizable Prompt Tuning for Audio-Language Models via Semantic Expansion

    Jaehyuk Jang*, W. Lee*, K. Ko*, Changick Kim

    Findings of the Annual Meeting of the Association for Computational Linguistics (ACL Findings), 2026

    [Paper] [arXiv]

    Generalizable Prompt Tuning for Audio-Language Models via Semantic Expansion

    Jaehyuk Jang*, W. Lee*, K. Ko*, Changick Kim

    Findings of the Annual Meeting of the Association for Computational Linguistics (ACL Findings), 2026

    [Paper] [arXiv]

    T-VSS: Test-Time Visual Subspace Steering for Adversarial Robustness of Vision-Language Models

    Jaehyuk Jang, M. Seo, S. Cho, K. Ko, Changick Kim

    arXiv, 2026

    [arXiv]

    T-VSS: Test-Time Visual Subspace Steering for Adversarial Robustness of Vision-Language Models

    Jaehyuk Jang, M. Seo, S. Cho, K. Ko, Changick Kim

    arXiv, 2026

    [arXiv]

    Efficient Test-Time Optimization for Depth Completion via Low-Rank Decoder Adaptation

    M. Seo*, W. Lee*, Jaehyuk Jang, Changick Kim

    arXiv, 2026

    [arXiv] [Project page]

    Efficient Test-Time Optimization for Depth Completion via Low-Rank Decoder Adaptation

    M. Seo*, W. Lee*, Jaehyuk Jang, Changick Kim

    arXiv, 2026

    [arXiv] [Project page]

    2025
    Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models

    S. Woo*, D. Kim*, Jaehyuk Jang*, Y. Choi, Changick Kim

    Findings of the Annual Meeting of the Association for Computational Linguistics (ACL Findings), 2025

    [Paper] [arXiv] [Project page]

    Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models

    S. Woo*, D. Kim*, Jaehyuk Jang*, Y. Choi, Changick Kim

    Findings of the Annual Meeting of the Association for Computational Linguistics (ACL Findings), 2025

    [Paper] [arXiv] [Project page]

    FCGNet: Foreground and Class Guided Network for Human Parsing

    Jaehyuk Jang, Y. Wang, Changick Kim

    Pattern Recognition (PR), 2025

    [Paper]

    FCGNet: Foreground and Class Guided Network for Human Parsing

    Jaehyuk Jang, Y. Wang, Changick Kim

    Pattern Recognition (PR), 2025

    [Paper]

    2024
    RITUAL: Random Image Transformations as a Universal Anti-hallucination Lever in LVLMs

    S. Woo*, Jaehyuk Jang*, D. Kim*, Y. Choi, Changick Kim

    arXiv, 2024

    [arXiv] [Project page]

    RITUAL: Random Image Transformations as a Universal Anti-hallucination Lever in LVLMs

    S. Woo*, Jaehyuk Jang*, D. Kim*, Y. Choi, Changick Kim

    arXiv, 2024

    [arXiv] [Project page]

    Towards Robust Multimodal Prompting With Missing Modalities

    Jaehyuk Jang, Y. Wang, Changick Kim

    IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

    [Paper] [arXiv]

    Towards Robust Multimodal Prompting With Missing Modalities

    Jaehyuk Jang, Y. Wang, Changick Kim

    IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

    [Paper] [arXiv]

    Subdivided Mask Dispersion Framework for Semi-Supervised Semantic Segmentation

    Y. Wang, Jaehyuk Jang, Changick Kim

    Pattern Recognition Letters (PRL), 2023

    [Paper]

    Subdivided Mask Dispersion Framework for Semi-Supervised Semantic Segmentation

    Y. Wang, Jaehyuk Jang, Changick Kim

    Pattern Recognition Letters (PRL), 2023

    [Paper]

    2023
    Noise-Augmented Missing Modality Aware Prompt based Learning for Robust Visual Recognition

    Y. Wang, Jaehyuk Jang, Changick Kim

    IEEE International Conference on Visual Communications and Image Processing (VCIP), 2023

    [Paper]

    Noise-Augmented Missing Modality Aware Prompt based Learning for Robust Visual Recognition

    Y. Wang, Jaehyuk Jang, Changick Kim

    IEEE International Conference on Visual Communications and Image Processing (VCIP), 2023

    [Paper]

    2021
    ChangeSim: Towards End-to-End Online Scene Change Detection in Industrial Indoor Environments

    J.-M. Park*, Jaehyuk Jang*, S.-M. Yoo, S.-K. Lee, U.-H. Kim, Jong-Hwan Kim

    IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021

    [Paper] [arXiv] [Project page] [NVIDIA Enterprise Supported]

    ChangeSim: Towards End-to-End Online Scene Change Detection in Industrial Indoor Environments

    J.-M. Park*, Jaehyuk Jang*, S.-M. Yoo, S.-K. Lee, U.-H. Kim, Jong-Hwan Kim

    IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021

    [Paper] [arXiv] [Project page] [NVIDIA Enterprise Supported]