Pengkun Jiao

Fudan University. Shanghai.

jpk.jpg

I am Pengkun Jiao (焦鹏昆), currently pursuing my doctoral studies at the School of Computer Science, Fudan University, under the supervision of Prof. Jingjing Chen and Prof. Yu-Gang Jiang. Additionally, I am also guided by Prof. Na Zhao and Prof. Bin Zhu.

My research interests focused on: LLMs, Multi-modal, Reinforce Learning, 3D.

💼 Here is my CV EN/中文.

news

Jun 26, 2025 One paper about multi-modal large language models is accepted by ICCV 2025.
Jan 20, 2025 Gave a talk at ECNU, “Visual Instruction Tuning for Downstream Tasks”
Oct 20, 2024 One paper about open-set single-source domain generalization (DEBUG) is accepted by TMM 2024.
Jul 11, 2024 My new website is online.
Jul 01, 2024 One paper about open-vocabulary 3D object detection (INHA) is accepted by ECCV 2024.

selected publications

  1. gas_eraser.jpg
    Don’t Deceive Me: Mitigating Gaslighting through Attention Reallocation in LMMs
    Pengkun Jiao, Bin Zhu, Jingjing Chen, and 2 more authors
    https://arxiv.org/abs/2504.09456, 2025
    HTML
  2. dual_lora.jpg
    From Holistic to Localized: Local Enhanced Adapters for Efficient Visual Instruction Fine-Tuning
    Pengkun Jiao, Bin Zhu, Jingjing Chen, and 2 more authors
    ICCV, 2025
    HTML
  3. rode.jpg
    RoDE: Linear Rectified Mixture of Diverse Experts for Food Large Multi-Modal Models
    Pengkun Jiao, Xinlan Wu, Bin Zhu, and 3 more authors
    arXiv preprint arXiv:2407.12730, 2024
    HTML
  4. inha.jpg
    Unlocking Textual and Visual Wisdom: Open-Vocabulary 3D Object Detection Enhanced by Comprehensive Guidance from Text and Image
    Pengkun Jiao, Na Zhao, Jingjing Chen, and 1 more author
    ECCV, 2024
    HTML
  5. debug.jpg
    Domain Expansion and Boundary Growth for Open-Set Single-Source Domain Generalization
    Pengkun Jiao, Na Zhao, Jingjing Chen, and 1 more author
    TMM, 2024
    HTML