avatar

Daeun Lee

Ph.D. Student
UNC Chapel Hill
daeun@cs.unc.edu


About Me

Thanks for visiting my website! ✨ I am a 2nd year Ph.D. student at UNC Chapel Hill, MURGe-Lab working with Prof. Mohit Bansal. Previously, I was advised by Prof. Jinkyu Kim at Korea University, Prof. Sung Ju Hwang at KAIST. I have also spent time as a Research Intern at Adobe Research (2025) and Meta (2026).

Research Direction

I believe technology should ultimately help people. Now, I am interested in how we can make a better world for underrepresented people through Multimodal AI.

My research interests include:

💗 If you are working on fun ideas related to video, AR glasses, or egocentric perception and reasoning, I’d love to collaborate!

News

2026
Jun VisionCoach is accepted to ECCV 2026 🇸🇪
May Join Meta Superintelligence Labs for 2026 Summer Research Intern!
Apr VideoRepair is accepted to ACL 2026 Findings 🌊
Mar New preprint is out: check out VisionCoach
Feb StreamGaze is accepted to CVPR 2026 🌅
2025
Dec New preprint is out: check out StreamGaze and IndustryNav
Aug Video-Skill-CoT is accepted to EMNLP 2025 Findings 🇨🇳
Jun New preprint is out: check out Video-Skill-CoT
May Join Adobe Research for 2025 Summer Research Intern!
Jan Invited talk at Cisco Meraki
2024
Nov New preprint is out: check out VideoRepair
Aug Started PhD journey at the UNC Chapel Hill MURGe-Lab 🎓
May BECoTTA is accepted to ICML 2024 🇦🇹
Apr One paper is accepted to CVPRW 2024 🇺🇸
2023
Jan One paper is oral-presented at WACV 2023 🏝️
2022
Dec I finished my first research internship at NAVER LABS!

Publications

  1. 🔎 Can robots reason about space in dynamic industrial environments?
    IndustryNav: Exploring Spatial Reasoning of Embodied Agents in Dynamic Industrial Navigation
    Yifan Li, Lichi Li, Anh Dao, Xinyu Zhou, Yicheng Qiao, Zheda Mai, Daeun Lee, Zichen Chen, Zhen Tan, Mohit Bansal, Yu Kong
    preprint
    PDF Code

  2. 🔎 Can grounded video reasoning models self-improve via visual prompting?
    ECCV
    VisionCoach: Reinforcing Grounded Video Reasoning via Visual-Perception Prompting
    Daeun Lee, Shoubin Yu, Yue Zhang, Mohit Bansal
    European Conference on Computer Vision (ECCV), 2026.
    PDF Code

  3. 🔎 Can text-to-video models detect and self-repair their own errors?
    ACL Findings
    VideoRepair: Self-Correcting Text-to-Video Generation with Misalignment Detection and Localized Refinement
    Daeun Lee, Jaehong Yoon, Jaemin Cho, Mohit Bansal
    The 64th Annual Meeting of the Association for Computational Linguistics (ACL) Findings, 2026.
    PDF Code

  4. 🔎 How far current streaming video understanding to AR glasses?
    CVPR
    StreamGaze: Gaze-Guided Temporal Reasoning and Proactive Understanding in Streaming Videos
    Daeun Lee, Subhojyoti Mukherjee, Branislav Kveton, Ryan A. Rossi, Viet Dac Lai, Seunghyun Yoon, Trung Bui, Franck Dernoncourt, Mohit Bansal
    The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026.
    PDF Code Accepted with strong scores (5/5/5) but desk-rejected after camera-ready because of coauthor's irresponsible reviewing.

  5. 🔎 Can skill-based reasoning help video models adapt across domains?
    EMNLP Findings
    Video-Skill-CoT: Skill-based Chain-of-Thoughts for Domain-Adaptive Video Reasoning
    Daeun Lee*, Jaehong Yoon*, Jaemin Cho, Mohit Bansal
    Empirical Methods in Natural Language Processing (EMNLP) Findings, 2025.
    PDF Code

  6. 🔎 Let's use mixture-of-domain experts for efficient continual test-time adaptation!
    ICML
    BECoTTA: Input-dependent Online Blending of Experts for Continual Test-time Adaptation
    Daeun Lee*, Jaehong Yoon*, Sung Ju Hwang
    International Conference on Machine Learning (ICML), 2024.
    PDF Code

  7. 🔎 Can HD maps teach lane detectors to generalize across domains?
    CVPRW
    HD Maps are Lane Detection Generalizers: A Novel Generative Framework for Single-Source Domain Generalization
    Daeun Lee, Minhyeok Heo, Jiwon Kim
    CVPR Data-Driven Autonomous Driving Simulation Workshop (CVPRW), 2024.
    PDF

  8. 🔎 How to detect rare objects in LiDAR point clouds?
    WACV
    Resolving Class Imbalance for LiDAR-based Object Detector by Dynamic Weight Average and Contextual Ground Truth Sampling
    Daeun Lee, Jinkyu Kim
    IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023.
    PDF

  9. 🔎 Can automatic colorization generalize beyond its training domain?
    ECCV
    Bridging the Domain Gap towards Generalization in Automatic Colorization
    Hyejin Lee, Daehee Kim, Daeun Lee, Jinkyu Kim, Jaekoo Lee
    European Conference on Computer Vision (ECCV), 2022.
    PDF

Services

Misc

🧸 I am a big fan of Sanrio (Hello Kitty, Mymelody, Kuromi), all teddy bears, and K-POP!
💐 Always appreciate all of my great advisors and collaborators!


Powered by Jekyll and Minimal Light theme.