Yoojin Oh

Yoojin Oh

Graduate student at KAIST

KAIST AI

About

Hi! I'm an M.S. student at KAIST AI, advised by Prof. Jong Chul Ye. I earned my B.S. in Artificial Intelligence from Ewha Womans University in February 2026 (Summa Cum Laude), where I was fortunate to be advised by Prof. Junhyug Noh throughout my undergraduate research.

My research lies at the intersection of generative modeling and robotics. I am interested in making robot foundation models faster and smaller for real-time deployment, with the ultimate goal of building robots that can work alongside people in everyday life.

News

Our paper FastOPD was released on arXiv.
I started my M.S. at the KAIST Graduate School of AI (Advisor: Prof. Jong Chul Ye).
Our paper Beyond Softmax was accepted to BMVC 2025.
Our paper SteeringTTA was accepted to the ICML 2025 PUT Workshop.

Publications

* Equal contribution

Comparison between DiffusionOPD, which queries the teacher at every denoising step, and FastOPD, which queries it at a single state reached by a flow-map jump

FastOPD: On-Policy Distillation for Lightweight VLA Deployment

Yoojin Oh*, Jeongsol Kim*, Yeonwoo Seo, Jangho Park, Seonghyun Jin, Sunwoo Park, Youngmin Kim, Youngjun Jun, Kyumin Choi, Jong Chul Ye

arXiv preprint, 2026

Beyond Softmax dual-branch sigmoid architecture overview

Beyond Softmax: Dual-Branch Sigmoid Architecture for Accurate Class Activation Maps

Yoojin Oh, Junhyug Noh

BMVC, 2025

SteeringTTA diffusion guidance method and qualitative results

SteeringTTA: Guiding Diffusion Trajectories for Robust Test-Time-Adaptation

Jihyun Yu, Yoojin Oh, Wonho Bae, Mingyu Kim, Junhyug Noh

ICML PUT Workshop, 2025

Education

Korea Advanced Institute of Science and Technology (KAIST)

M.S. in Artificial Intelligence · GPA: 4.3/4.3

BISPL @ KAIST AI — BioImaging, Signal Processing, & Machine Learning Lab

EPITA: École pour l'informatique et les techniques avancées

Artificial Intelligence Summer Program · GPA 4.0/4.0

Ewha Womans University

B.S. in Artificial Intelligence · GPA: 4.23/4.3 · Summa Cum Laude

Experience

Undergraduate Research Intern

Computer Vision Lab, POSTECH · Advisor: Suha Kwak

  • Self-supervised audio-visual multimodal learning for multi-source sound localization using slot attention.
  • Foundations of vision-language-action (VLA) models.

Undergraduate Research Intern

Practical AI Lab, Ewha Womans University · Advisor: Junhyug Noh

  • Test-time adaptation based on diffusion models for robust learning.
  • Improve class activation maps (CAMs) for weakly supervised object localization.

Research Intern

SPRESTO Corp.

  • Development of a spycam detection model to help prevent illegal filming.

Industry Project

NAVER Z (ZEPETO)

  • Fine-tune generative models to create virtual objects for ZEPETO, a metaverse platform.

Honors & Awards

Kim Aeda Award

Ewha Womans University · Awarded to the graduating student with the best GPA in each college.

3rd Place, 2024 SPARCS Service Hackathon

Korea Advanced Institute of Science and Technology (KAIST)

1st Place, 2023 College of AI Ideas Competition

Ewha Womans University

National Scholarship for Outstanding Students in Science and Engineering

Korea Student Aid Foundation (KOSAF)

Scholarship for Academic Excellence

Ewha Womans University · Awarded three times

Song Bok Eun Scholarship

Ewha Womans University

Kim Hyun Gon Shin Kyung Soon Scholarship

Ewha Womans University

Hwain Chang Lee Scholarship

Ewha Womans University

Kathryne B. Sears Memorial Scholarship

Ewha Womans University