Myungbeom Her

Hello! I’m a Ph.D. candidate at DGIST advised by Prof. Jinhyun So.

I am working on how machines can train effectively in edge device. My research centers on gradient-free optimization Algorithm and On-device AI.

Ultimately, I aim to develop intelligent systems that can learn efficiently in real time on edge devices by leveraging zeroth-order optimization and local learning to reduce activation memory and training time.

Current: Ph.D. @ DGIST (CORE-AI Lab)

Contact: myungbeom.her@dgist.ac.kr

github: https://github.com/MyungBeomHer

google scholar: https://scholar.google.com

PUBLICATIONS

Publised

FRU-Adapter: Frame Recalibration Unit Adapter for Dynamic Facial Expression Recognition

Myungbeom Her, Hamza Ghulam Nabi, and Ji-Hyeong Han*

Electronics 2025

Paper, code

Batch Transformer: Look for attention in batch

Myungbeom Her, Jisu Jeong, Hojoon Song, and Ji-Hyeong Han*

Paper code

IEEE ACCESS 2025

Submit

PTN: Generating a learnable noise in a neural network to enhance robustness to occlusion

Myungbeom Her, Ji-Hyeong Han*

code

Image and Vision Computing 2026

AWARDS AND HONORS

Future Talent Scholarship (2023.03 ~ 2024.12)

Seoul National University of Science and Technology

Department autonomous scholarship (2024.11)

Seoul National University of Science and Technology

1st Prize (2022.11)

Artificial potential-based planner algorithm competition

Yonsei University

Excellence Prize (2022.08)

SW Startup Idea Contest

Yonsei University

3th Prize (2022.06)

Programming Competition

Yonsei University

1st Prize (2021.12)

AI Competition

Yonsei University

Research Experience

DGIST (2026.02 ~ present)

Ph.D. student at CORE-AI Lab, supervised by Prof. Jinhyun So

Seoul National University of Science and Technology (2023.03 ~ 2025.02)

M.S. student at HCIR Lab, supervised by Prof. Ji-Hyeong Han

Yonsei University (2022.09 ~ 2022.11)

Undergraduate Industry-academia student at 선율.Inc, supervised by Prof. Park jun mo

Page updated

Google Sites

Report abuse