cv
General information
| Name | Kei Hiroshima |
| Label | PhD Student |
| hiroshima-kei-st@ynu.jp | |
| Url | https://KeiHiroshima.github.io |
| Summary | I study how machine learning systems can adapt and optimize efficiently in real-world settings, where data arrives sequentially in non-stationary streams and problems are underspecified. My work spans preference-aware model merging for continual learning and LLM-driven auto-formulation and algorithm selection for black-box optimization. I am broadly interested in continual learning and closing the gap between machine learning and optimization. |
Work
-
2025.04 - 2026.03 Part-time Lecturer
Yokohama National University
- Course: Introduction to Mathematics, Data Science and AI.
Education
-
2025.04 - Present Yokohama, Japan
Ph.D. Student
Graduate School of Environment and Information Sciences, Yokohama National University
Shirakawa/Uchida Laboratory
- Expected graduation: spring 2028.
-
2023.01 - 2023.12 Sydney, Australia
Exchange Program
Faculty of Engineering and Information Technology, University of Technology Sydney
Exchange Student
-
2022.04 - 2025.03 Yokohama, Japan
Master of Informatics
Graduate School of Environment and Information Sciences, Yokohama National University
Informatics
-
2018.04 - 2022.03 Yokohama, Japan
Bachelor of Engineering
Department of Mathematics, Physics, Electrical Engineering and Computer Science, College of Engineering Science, Yokohama National University
Electrical and Computer Engineering Program
Awards
- 2026
Best Paper Award
FAN2026 (The 34th Intelligent System Symposium)
Evaluating Large Language Models on Black-Box Optimization Word Problems
- 2025
Outstanding Presentation Award
SSI 2025 (SICE Symposium on Systems and Information)
Skills
| Research Areas | |
| Artificial Intelligence | |
| Machine Learning | |
| Continual Learning | |
| Transfer Learning | |
| Black-box Optimization | |
| Natural Language Processing |
Languages
| Japanese | |
| Native speaker |
| English | |
| TOEIC 860 / IELTS 6.5 |