avatar

Phillip Y. Lee

이유승

PhD Student

KAIST Graduate School of AI

About Me

I am Phillip Lee, a second year PhD student at KAIST Graduate School of AI, advised by Prof. Minhyuk Sung. I am also a student researcher at Google DeepMind, Mountain View. I received both my Masters and Bachelors degree at KAIST. I am also fortunate to closely collaborate with Prof. Leonidas Guibas. I am a recipient of Qualcomm Innovation Fellowship Korea 2025.

My research interests revolve around analyzing and enhancing the spatial understanding of multimodal foundation models, with the broader goal of building truly autonomous and embodied agents capable of operating in the physical world. Specifically, I have worked on advancing the 3D spatial reasoning capabilities of vision-language models (VLMs), and enhancing the spatial control of diffusion models.

I am open to collaboration opportunities! Please feel free to contact me via email. My CV is at Curriculum Vitae (CV).

Work Experience

News

[ see more ]

Publications

Selected publications.

  1. Hidden Sensitivity in Spatial Reasoning Evaluation: Diagnosis and Re-ranking with VSI-Bench
    Phillip Y. Lee, Jin Yoo, Minseo Kim, Leonidas Guibas , Minhyuk Sung
    Combining Theory and Benchmark (CTB) Workshop at ICML 2026
    PDF
    Token Warping Helps MLLMs Look from Nearby Viewpoints
    Phillip Y. Lee*, Chanho Park*, Mingue Park, Seungwoo Yoo, Juil Koo, Minhyuk Sung (* equal contribution)
    CVPR 2026
    Project arXiv
    Toward Ambulatory Vision: Learning Visually-Grounded Active View Selection
    Juil Koo*, Daehyeon Choi*, Sangwoo Youn*, Phillip Y. Lee, Minhyuk Sung (* equal contribution)
    Reinforcement Learning from World Feedback (RLxF) Workshop at ICML 2026
    Project PDF arXiv
    DiverseVAR: Balancing Diversity and Quality of Next-Scale Visual Autoregressive Models
    Mingue Park*, Prin Phunyaphibarn*, Phillip Y. Lee, Minhyuk Sung
    ECCV 2026
    arXiv
    Unconditional Priors Matter! Improving Conditional Generation of Fine-Tuned Diffusion Models
    Prin Phunyaphibarn, Phillip Y. Lee, Jaihoon Kim, Minhyuk Sung
    WACV 2026
    Project PDF arXiv
    Perspective-Aware Reasoning in Vision-Language Models via Mental Imagery Simulation
    Phillip Y. Lee, Jihyeon Je, Chanho Park, Mikaela Angelina Uy, Leonidas Guibas, Minhyuk Sung
    ICCV 2025 / Human to Robot (H2R) Workshop at CoRL 2025
    Project PDF arXiv
    GrounDiT: Grounding Diffusion Transformers via Noisy Patch Transplantation
    Phillip Y. Lee*, Taehoon Yoon*, Minhyuk Sung (* equal contribution)
    NeurIPS 2024
    Project PDF arXiv
    ReGround: Improving Textual and Spatial Grounding at No Cost
    Phillip Y. Lee, Minhyuk Sung
    ECCV 2024
    Project PDF arXiv BibTex Slides
    SyncDiffusion: Coherent Montage via Synchronized Joint Diffusions
    Phillip Y. Lee, Kunho Kim, Hyunjin Kim, Minhyuk Sung
    NeurIPS 2023
    Project PDF arXiv Code BibTex Demo Talk

Teaching

Academic Services

Achievements

Powered by Jekyll and Minimal Light theme.