Jiyun Kim
Jiyun Kim / κΉ€μ§€μœ€
Researcher
JiyunKim@kaist.ac.kr
0009-0003-0026-4495
Linkedin
jixyuxn
12889385

Description

Jiyun Kim received her B.S. degree in Robot Engineering from Keimyung University, South Korea, in 2024, and her M.S. degree in Artificial Intelligence from the Gwangju Institute of Science and Technology (GIST), South Korea. She is currently conducting research at the Korea Advanced Institute of Science and Technology (KAIST).

Her research interests include learning-based motion planning, autonomous driving, and intelligent robotic systems. She focuses on developing reliable and robust decision-making methods based on imitation learning, vision-language models and vision-language-action models (VLMs/VLAs), and robot foundation models. Her long-term goal is to build safer and more human-centered autonomous systems that can operate effectively in complex real-world environments.

Publications


Quick Links


Papers Under Review


    2026
  1. [New] CAR Planner: Constrained-Attention-Based Robust Imitation Learning for Autonomous Driving
    Jiyun Kim , Kyunghwan Choi*
    Withheld during double-blind review , 2026 [πŸ“ƒ Full-Text]

International Conference Papers


    2026
  1. [New] TRIAD Planner: Three-Level Robust Imitation Learning for Autonomous Driving
    Jiyun Kim , Ahmad Mouri Zadeh Khaki, Kyunghwan Choi*
    Conference on Robot Learning (CoRL) (accepted, in press), 2026 [πŸ“ƒ Full-Text]
  2. CARE Planner : Constrained Attention and Risk-aware Planning for Imitation based Autonomous Driving
    Jiyun Kim , Kyunghwan Choi*
    International Federation of Automatic Control (IFAC) World Congress (accepted, in press), 2026 [πŸ“ƒ Full-Text]

    3. 2025
  3. Mitigating Attention Collapse via Mean-Deviation Constrained Optimization
    Jiyun Kim , Kyunghwan Choi*
    International Conference on Robot Intelligence Technology and Applications (RiTA) (accepted, in press), 2025 [🌐 Preprint] [πŸ“ƒ Full-Text]
      πŸ“„ Awarded Best Intelligent Paper Award at the 2025 International Conference on Robot Intelligence Technology and Applications (RiTA).
  4. Imposing a Weight Norm Constraint for Neuro-Adaptive Control
    Myeongseok Ryu, Jiyun Kim , Kyunghwan Choi*
    European Control Conference (ECC) , pp. 380-385, 2025 [🌐 Online] [πŸ“ƒ Full-Text] [πŸŽ₯ Presentation] [πŸ’Ύ Cite] [GitHub logo Code]

Domestic Conference Papers


    2026
  1. CARE Planner : Constrained Attention and Risk-aware Planning for Imitation-based Autonomous Driving
    CARE Planner : λͺ¨λ°©ν•™μŠ΅ 기반 μžμœ¨μ£Όν–‰μ„ μœ„ν•œ μ œμ•½ μ–΄ν…μ…˜ 및 μœ„ν—˜ 인지 ν”Œλž˜λ‹
    Jiyun Kim , Kyunghwan Choi*
    μ œμ–΄λ‘œλ΄‡μ‹œμŠ€ν…œν•™νšŒ (ICROS) (accepted, in press), 2026 [πŸ“ƒ Full-Text]
      πŸ“„ Awarded Best Paper Award at the 2026 Institute of Control, Robotics and Systems (ICROS) Conference.

    2. 2025
  2. CAR Planner: Constrained-Attention-Based Robust Imitation Learning for Autonomous Driving
    CAR Planner: μ œμ•½ μ–΄ν…μ…˜ 기반 강건 λͺ¨λ°©ν•™μŠ΅ ν”Œλž˜λ„ˆ
    Jiyun Kim , Kyunghwan Choi*
    ν•œκ΅­μžλ™μ°¨κ³΅ν•™νšŒ μΆ”κ³„ν•™μˆ λŒ€νšŒ (KSAE) (accepted, in press), 2025 [πŸ“ƒ Full-Text]

Master's Theses


    2026
  1. Robust Imitation Learning with Attention Constraints and Risk Awareness for Motion Planning in Autonomous Driving
    Jiyun Kim , Kyunghwan Choi*
    Gwangju Institute of Science and Technology (GIST) Library, 2026 [πŸ“ƒ Full-Text]

Preprints


    2025
  1. CAR Planner: Constrained-Attention-Based Robust Imitation Learning for Autonomous Driving
    Jiyun Kim , Kyunghwan Choi*
    Techrxiv, 2025 [🌐 Preprint Online] [πŸ“ƒ Preprint Full-Text]