|
Kyungho Bae (배경호) I'm a machine learning engineer at Karrot(daangn). I received my M.S. in Artificial Intelligence from Kyung Hee University, where I worked under the supervision of Prof. Jinwoo Choi. During my graduate studies, I also worked as a research intern at LG AI Research. My research interests focus on video understanding, multimodal large language models, and representation learning, with an emphasis on building robust and generalizable systems for complex real-world environments. |
|
News2025
2024
2023
|
Employment
|
ResearchI work on representation learning for video and multimodal understanding, with interests in domain adaptation, continual learning, and reducing bias in learned models. The following are my recent publications. († indicates the first authors, * indicates the corresponding authors) |
|
ESSENTIAL: Episodic and Semantic Memory Integration for Video Class-Incremental Learning Jongseo Lee†, Kyungho Bae†, Kyle Min, Gyeong-Moon Park*, Jinwoo Choi* ICCV 2025 Keywords: Continual Learning, Video Action Recognition arXiv coming soon |
|
MASH-VLM: Mitigating Action-Scene Hallucination in Video-LLMs through Disentangled Spatial-Temporal Representations Kyungho Bae, Jinhyung Kim, Sihaeng Lee, Soonyoung Lee, Gunhee Lee*, Jinwoo Choi* CVPR 2025 (Highlight, acceptance rate = 3.71%) Keywords: Hallucination, Video-LLM, Multimodal arXiv |
|
DEVIAS: Learning Disentangled Video Representations of Action and Scene for Holistic Video Understanding Kyungho Bae†, Geo Ahn†, Youngrae Kim†, Jinwoo Choi* ECCV 2024 (Oral, acceptance rate = 2.33%) Keywords: Video Representation, Bias in Video arXiv / code |
|
GLAD: Global-Local View Alignment and Background Debiasing for Unsupervised Video Domain Adaptation Hyogun Lee†, Kyungho Bae†, Yumin Ko, Seongjong Ha, Gyeongmoon Park*, Jinwoo Choi* WACV 2024 Keywords: Video Domain Adaptation arXiv / code |
Education
|
|
Based on Jon Barron's website. Modified by Kyungho Bae. |