Hyunah Ko

I am an M.S. student at the Kim Jaechul Graduate School of AI, KAIST, working in CVLAB, advised by Prof. Seungryong Kim. I received my B.S. in Civil and Environmental Engineering with a double major in Computer Science from Yonsei University. Previously, I worked with Prof. Seong Jae Hwang as an undergraduate researcher.

Email CV Scholar Github LinkedIn

drag the background to orbit · click to reconstruct

Education

KAIST, Kim Jaechul Graduate School of AI · Sep. 2026 – Present
M.S. student, CVLAB / Advisor: Prof. Seungryong Kim

Yonsei University · Mar. 2022 – Aug. 2026
B.S. in Civil and Environmental Engineering, Double Major: Computer Science / GPA: 3.96 / 4.5

Research Interests

My research focuses on visual representation learning for 3D/4D understanding and robotics. I am interested in multi-view learning and spatial representations that capture the geometry and semantics of the physical world. My interests also include compact 3D/4D reconstruction and world action models (WAMs), with the long-term goal of connecting visual perception with action.

Multi-view representation learning and 3D scene understanding
Compact 3D/4D representations and feed-forward scene reconstruction
World action models and visual representations for robotics

Publications

Publications and preprints in 3D/4D computer vision and robotics. I also publish as Hyuna Ko. (* denotes equal contribution) See CV for the complete list.

	One Loss Is Not Enough: Cross-Objective Structured Pruning for Reasoning Vision–Language–Action Models Seongchan Kim, Soowon Son, Sangoh Kim, Hyunah Ko, Eunbeen Hong, Hyunkoo Lee, Sanghyun Lee, WonJun Moon, Kihong Kim, Dongyeon Son, Woongchan Nam, Seungryong Kim NeurIPS Workshops, 2026 Grounded and Faithful Vision-Language Models for Real-World Deployment · On-Device Intelligence: Foundation Models under Real-World Constraints · PTA: From Pretrained Representations to Acting Agents
	C4G: Learning Global Motion with Compact Gaussians for Feed-Forward 4D Reconstruction Mungyeom Kim, Minkyeong Jeon, Honggyu An, Jaewoo Jung, Hyuna Ko, Jisang Han, Hyeonseo Yu, Donghwan Shin, Sunghwan Hong, Takuya Narihira, Kazumi Fukuda, Yuki Mitsufuji, Seungryong Kim arXiv preprint*, 2026 arXiv / project page
	C3G: Learning Compact 3D Representations with 2K Gaussians Honggyu An, Jaewoo Jung, Mungyeom Kim, Chaehyun Kim, Minkyeong Jeon, Jisang Han, Kazumi Fukuda, Takuya Narihira, Hyuna Ko, Junsu Kim, Sunghwan Hong, Yuki Mitsufuji, Seungryong Kim CVPR, 2026 Also at the CVPR 2026 Workshop on Open-World 3D Scene Understanding and Representations arXiv / project page
	Backbone Augmented Training for Adaptations Jae Wan Park, Junhyeok Kim, Youngjun Jun, Hyunah Ko, Seong Jae Hwang AAAI 2026 Workshop on Artificial Intelligence with Biased or Scarce Data — Oral arXiv preprint, 2025 arXiv Improving model adaptation performance through backbone augmented training techniques.
	Probing and Leveraging Video Diffusion Transformer Features for Robust Point Tracking Soowon Son, Honggyu An, Jisu Nam, Hyunah Ko, Chaehyun Kim, Dahyun Chung, Siyoon Jin, Jung Yi, Junhwa Hur, Seungryong Kim arXiv preprint, 2025; revised 2026 · DiTracker arXiv / project page
	Fourier Decomposition for Explicit Representation of 3D Point Cloud Attributes Donghyun Kim, Chanyoung Kim, Hyunah Ko, Seong Jae Hwang arXiv preprint, 2025; revised 2026 arXiv Proposing a novel method for explicit representation of 3D point cloud attributes using Fourier decomposition.

Research Experience

Research
Experience

CVLAB, KAIST (Jul. 2025 - Present)
M.S. student (since Sep. 2026); previously Undergraduate Researcher. Advisor: Prof. Seungryong Kim

Medical Imaging and Computer Vision Lab, Yonsei Univ. (Jul. 2024 - Jul. 2025)
Undergraduate Researcher, advised by Prof. Seong Jae Hwang

Environmental Biotechnology & Thermal Lab, Yonsei Univ. (Jul. 2023 - Sep. 2023)
Undergraduate Researcher, advised by Prof. Joonhong Park

Selected Projects

	Multi-View Diffusion Gaussian Splatting (MVDGS) Apr. 2025 – May. 2025 \| Team Leader code Team Project, Yonsei Artificial Intelligence Developed a single-image-to-3D generation framework integrating CameraCtrl and 3D Gaussian Splatting for temporally and geometrically consistent novel view synthesis. Awarded 3rd Place at the 6th YAICON.
	Recipe Recommendation via Object Detection and Knowledge Graph QA Oct. 2024 – Nov. 2024 \| Team Leader code Team Project, Yonsei Artificial Intelligence Built an end-to-end recipe recommendation system that detects ingredients from images using YOLOv5 and generates personalized recipes with knowledge-augmented LLM prompts and Stable Diffusion food images.
	Generative Pet Memorials with LLaMA3 and Stable Diffusion Aug. 2024 – Sep. 2024 code Team Project, Yonsei Data Science Lab Created a pet memorial application that generates personalized letters with LoRA-finetuned LLaMA3 and commemorative artwork with Stable Diffusion, ControlNet, and style transfer.
	Seoul Night Bus Route Analysis and Design Jul. 2024 – Aug. 2024 \| Team Leader report Team Project, Yonsei Data Science Lab Analyzed Seoul's late-night mobility and night bus ridership data to identify underserved regions, and proposed data-driven night bus routes using exploratory data analysis and regression-based demand modeling.

Extracurricular activities

Honors &
Awards

🥉 3rd Place, 6th YAICON, Yonsei Artificial Intelligence (2025)
Awarded for MVDGS, a single-image-to-3D generation framework combining video diffusion and 3D Gaussian Splatting.

🏆 Grand Prize, 2024 Jeju Satellite Data Utilization Competition (Jul. 2024 – Aug. 2024)
Proposed a geospatial solution to address harmful green algae blooms in Jeju Island, using a segmentation model on satellite imagery to detect and map the bloom distribution.

🎖️ 77th out of 514, AirQo African Air Quality Prediction Challenge, Zindi (Mar. 2024 – Jun. 2024)
Addressed the challenge of sparse ground-level sensors in African cities by developing a model to forecast hourly PM2.5 concentrations from satellite and meteorological data.

Leadership

Yonsei Artificial Intelligence (joined Dec. 2023)

Vice President (2025)

Yonsei Computer Club (joined Jul. 2023)

Vice President (Feb. 2024 - Feb. 2025)

Member, Yonsei Data Science Lab (Dec. 2023 - Nov. 2024)

Personal Life

Outside of research, I'm passionate about music, especially live band performances. I frequently attend concerts and enjoy discovering new artists. Music helps me stay creative and inspired in both my personal and academic life.