SeungBum Ha

Ph.D. Student @ UNIST

|

About

SeungBum Ha

I am a Ph.D. student at Ulsan National Institute of Science and Technology (UNIST), Graduate School of Artificial Intelligence. My research focuses on building trustworthy and efficient machine learning systems, particularly in machine unlearning and federated learning.

Outside the lab, I love staying active — you can find me catching waves 🏄, sending bouldering routes 🧗, or playing pickup basketball 🏀.

Always happy to connect — whether it's a research collaboration, an interesting opportunity, or just a casual coffee chat ☕.

Education

Ph.D. in Artificial Intelligence

UNIST · 2024 – Present · Advisor: Sung Whan Yoon, Saerom Park (co-advisor)

M.S. in Artificial Intelligence

UNIST · 2022 – 2024 · Advisor: Sung Whan Yoon

B.S. in Industrial Engineering

UNIST · 2019 – 2022 · Advisor: Sunghoon Lim

Magna Cum Laude · Early Graduation (7 semesters)

Publications

preprint

On-Trajectory Coding for Optimizing End-to-End Latency of Image Semantic Communications via Diffusion Models

J Park, SB Ha, SW Yoon

Under Review, 2026

preprint

Unlearning's Blind Spots: Over-Unlearning and Prototypical Relearning Attack

SB Ha, S Park, SW Yoon

arXiv preprint (Under Review), 2025

journal

Benchmarking Federated Learning for Semantic Datasets: Federated Scene Graph Generation

SB Ha*, T Lee*, J Lim, SW Yoon

Pattern Recognition Letters, Vol. 197, pp. 195-201, 2025

conference

MetaVers: Meta-Learned Versatile Representations for Personalized Federated Learning

JH Lim*, SB Ha*, SW Yoon

WACV 2024, pp. 2587-2596

Awards & Honors

🥈 ANTI-BAD Challenge, 2nd Place

4th IEEE Conference on Secure and Trustworthy Machine Learning · 2026

Graduate Full Scholarship

UNIST · 2022

UNIST, KAIST, POSTECH · 2021

Korea University · 2021

Undergraduate Full Scholarship

UNIST · 2019

Experience

Research Journey

Research Student

Machine Intelligence and Information Theory Lab

UNIST

Advisor: Sung Whan Yoon, Saerom Park (co-advisor)

2022.08 – Present

Undergraduate Research Intern

Homomorphic Encryption Lab

UNIST

Advisor: Miran Kim

2021.12 – 2022.07

Research Intern

Unstructured Data Mining and Machine Learning Lab

UNIST

Advisor: Sunghoon Lim

2020.05 – 2021.12

More

Patents

Meta-learning algorithms for personalized federated learning

Sung Whan Yoon, Jinhyuk Lim, SeungBum Ha

KR-Application No. 10-2023-0037350 · 2023

Academic Service

Reviewer

ICLR 2025, ICLR 2026, CVPR 2026, ECCV 2026, NeurIPS 2026

Teaching

Distributed Learning

Teaching Assistant, UNIST (AI51501)

Sep. 2024 – Dec. 2024

Introduction to AI Safety

Teaching Assistant, UNIST (IE47001)

Mar. 2024 – Jun. 2024

AI Novatus Academia Curriculum

Teaching Assistant, Gyeongsangnam-do

Sep. 2022 – Feb. 2023

Projects

Hierarchical federated learning simulations and clients contribution measurement

ETRI, Korea · Jun. 2024 – Nov. 2024

Multi-inference performance based federated learning asynchronous consensus

ETRI, Korea · Jun. 2023 – Nov. 2023

Development of feasibility verification technology for resource hidden deep learning model

ETRI, Korea · Apr. 2023 – Nov. 2023

Predicting stock price volatility using FinBERT

AICP Program, UNIST · Mar. 2021 – Dec. 2021

Manufacturing Bigdata Analysis — anomaly detection based on manufacturing process data

Hankuk Mold, Korea · Nov. 2020 – Feb. 2021

Energy Consumption Analysis for Additive Manufacturing

UNIST UDMML Lab · Jul. 2020 – Jul. 2021

Optimization of AI-based fusion manufacturing process conditions and prediction of defects

Seoyeon E-hwa / Inter-X, Korea · Jun. 2020 – Feb. 2021