Gwangyeol Yu

Gwangyeol Yu

M.S. Student

Korea University

rhkdduf627@korea.ac.kr

About

I am an M.S. student in the Trustworthy AI Lab at Korea University, advised by Prof. Jongheon Jeong. My research interests lie in trustworthy generative models and structured visual understanding, with a particular focus on the reliability and controllability of diffusion models and object-centric representation learning. Ultimately, I aim to build AI systems that operate reliably in real-world settings and can be trusted by people.

Publications

(C: Conference, J: Journal, P: Patent Application; *: Equal contribution)

Robust Finetuning from Generative Representations via Nuisance-extended Information Bottleneck thumbnail

[Preprint] Robust Finetuning from Generative Representations via Nuisance-extended Information Bottleneck

Minsun Lee, Gwangyeol Yu, Yunhun Nam, Jinwoo Shin, Jongheon Jeong

Under review.

Refining Visual Artifacts in Diffusion Models via Explainable AI-based Flaw Activation Maps thumbnail

[Preprint] Refining Visual Artifacts in Diffusion Models via Explainable AI-based Flaw Activation Maps

Seoyeon Lee*, Gwangyeol Yu*, Chaewon Kim*, Jonghyuk Park

Under review.

Mimic Blocker: Self-Supervised Adversarial Training for Voice Conversion Defense with Pretrained Feature Extractors thumbnail

[C1] Mimic Blocker: Self-Supervised Adversarial Training for Voice Conversion Defense with Pretrained Feature Extractors

Gwangyeol Yu*, Junhyuk Lee*, Seoryeong Kim*, Jimin Lee, Jehyuk Lee

Interspeech, 2025.

Distortion Measurement Metric for Human Image Refinement and Evaluation Using Distorted Image Datasets thumbnail

[J1] Distortion Measurement Metric for Human Image Refinement and Evaluation Using Distorted Image Datasets

Gwangyeol Yu*, Kyungseo Yoon*, Sangwoo Lee, Yeju Shin, Jonghyuk Park

IEEE Access, vol. 13, pp. 34390-34408, 2025.

Development and Contentization of Athlete Fitness Indicators Using STT and Object Tracking thumbnail

[P1] Development and Contentization of Athlete Fitness Indicators Using STT and Object Tracking

Gwangyeol Yu, Seoyeon Lee, Kiseoup Shin, Seunghak Lee

KR Patent Application No. 10-2024-0007425.

Education

Korea University

M.S. Student in Artificial Intelligence

Kookmin University

B.A. in AI, Bigdata & Management

Minor: School of Software | Interdisciplinary Major: Data Science

Awards & Honors

HD Hyundai AI Challenge - Grand Prize

Hyundai

2023 BigContest - Grand Prize

Ministry of Science and ICT, NIA

Hackathon Competition for the Disabled - Honorable Mention

Ministry of Health and Welfare

DACrew 4th - Encouragement Award

DACON

KAMP: The 2nd K-Artificial Intelligence Manufacturing Data Analysis Competition - Gold Prize

KAIST, KOSMO, Ministry of SMEs and Startups

2022 BigContest - Gold Prize

Ministry of Science and ICT, NIA

SKT AI Fellowship 4th - 1st, Encouragement Prize

SK Telecom

The 2nd NH Investment & Securities Big Data Competition - Honorable Mention

NH Investment & Securities

2021 BigContest - Encouragement Award

Ministry of Science and ICT, NIA