I am a Research Officer at the Agency for Defense Development, Korea's national defense R&D agency (akin to DARPA), and a First Lieutenant in the Republic of Korea Army. I received my B.S. from DGIST's interdisciplinary program, concentrating in Computer Science & Engineering.

Research Interests
My research lies at the intersection of computer vision and robotics, aiming to enable agents to generalize to novel scenarios and execute long-horizon tasks. I am currently developing unified representations that integrate dynamic 3D geometry with open-world semantics.
My research encompasses three key areas:
1. Spatiotemporal Scene Representations (Ongoing)
: Modeling scene dynamics from online observations
2. Vision-Language Representations [ECCV 2024]
: Learning open-world semantic representations that enable generalization
3. Implicit Neural Representations [CVPR 2023]
: Reconstructing continuous scene representations from sensor data

Publications

NeurIPS 2025

Tortoise and Hare Guidance: Accelerating Diffusion Model Inference with Multirate Integration

Yunghee Lee, Byeonghyun Pak, Junwha Hong, Hoseong Kim

ECCV 2024

Textual Query-Driven Mask Transformer for Domain Generalized Segmentation

Byeonghyun Pak*, Byeongju Woo*, Sunghwan Kim*, Dae-hwan Kim, Hoseong Kim

CVPR 2023

🏆 Highlight (Top 2.5%)

B-spline Texture Coefficients Estimator for Screen Content Image Super-Resolution

Byeonghyun Pak*, Jaewon Lee*, Kyong Hwan Jin

Ongoing Projects

Active

4D Neural Implicit Mapping

We build dense representations to jointly encode geometric structure and high-level semantics of dynamic environments.

In submission

Aligning Forest and Trees in Images and Long Captions for Visually Grounded Understanding

Byeongju Woo, Zilin Wang, Byeonghyun Pak, Sangwoo Mo, Stella X. Yu

We propose a hierarchical vision-language model that enables fine-grained multimodal understanding without extra annotations.

Experience

Research Officer for National Defense

Manager: Dr. Eunjin Koh/Advisor: Dr. Hoseong Kim
  • Developed object detection models for surveillance UAVs
  • Worked on domain generalization with vision-language models [ECCV 2024]
  • Worked on image & video diffusion models [NeurIPS 2025]

Undergraduate Research Intern

Advisor: Prof. Kyong Hwan Jin
  • Worked on implicit neural representations (INRs) for super-resolution [CVPR 2023]

Education

Daegu Gyeongbuk Institute of Science and Technology (DGIST)

B.S. in Engineering, School of Undergraduate StudiesFeb 2019 – Feb 2023

University of California, Berkeley (UCB)

Summer Visiting StudentJul 2019 – Aug 2019