Myungseo Song

I recently graduated from Seoul National University with a B.S. in Computer Science and Engineering. Previously, I worked as a machine learning researcher at AI startups for 3.5 years, in collaboration with Prof. Tae-Hyun Oh and Prof. Jong-Seok Lee. Before that, I worked as an undergraduate researcher in the Computer Vision Lab at Seoul National University under the supervision of Prof. Bohyung Han.

Email / Scholar / Github

Education

Seoul National University Mar. 2018 – Feb. 2026
B.S., Computer Science and Engineering
* Includes three-year mandatory military service in South Korea

Publications

	Exploring the Camera Bias of Person Re-identification Myungseo Song, Jin-Woo Park, Jong-Seok Lee ICLR 2025 (Spotlight paper) Paper / OpenReview
	A Training-Free Defense Framework for Robust Learned Image Compression Myungseo Song, Jinyoung Choi, Bohyung Han arXiv preprint, 2024 Paper
	Unsupervised Pre-training for Data-Efficient Text-to-Speech on Low Resource Languages Seongyeon Park, Myungseo Song, Bohyung Kim, Tae-Hyun Oh (Equal contribution) ICASSP 2023* * Also presented at ICML 2022 Workshop on Machine Learning for Audio Synthesis (Oral presentation) Paper / Code
	Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform Myungseo Song, Jinyoung Choi, Bohyung Han ICCV 2021 Paper / Code

Research Experience

mAy-I, Inc. Nov. 2023 – May. 2025
Lead Machine Learning Researcher Seoul, Korea
* Fulfilled mandatory military service

CNAI, Inc. Oct. 2021 – Oct. 2023
Machine Learning Researcher Seoul, Korea
* Fulfilled mandatory military service

Computer Vision Lab at Seoul National University Sep. 2020 – Sep. 2021
Undergraduate Researcher, Advisor: Prof. Bohyung Han Seoul, Korea

NCSOFT, Inc. Jul. 2020 – Aug. 2020
Research Intern Pangyo, Korea

Data Mining Lab at Seoul National University (UROP) Mar. 2020 – Jun. 2020
Undergraduate Researcher, Advisor: Prof. U Kang Seoul, Korea

Academic Service

Journal Reviewer
TIP (2023)

Conference Reviewer
ICLR (2025), NeurIPS (2024-2025), WACV (2023)

This website is built with Jon Barron's template.