Scholar
Junghyun Lee
Google Scholar ID: sYtSnnQAAAAJ
PhD Student @ KAIST AI
Citations & Impact
Citations
99
H-index
6
i10-index
5
Publications
20
Co-authors
18
list available
Publications
Resume (English only)
Academic Achievements
- Published several papers, including:
- 1. AdaSTaR: Adaptive Data Sampling for Training Self-Taught Reasoners
- 2. A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits
- 3. Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion
- 4. Nearly Optimal Latent State Decoding in Block MDPs
- Additionally, will serve as a reviewer for ICLR 2026 and AISTATS 2026, and one paper accepted to NeurIPS 2025.
Research Experience
- Was a Research Scientist Intern at Adobe Research, San Jose, during the summer of 2025 (May ~ Aug), working with Branislav “Brano” Kveton, Subhojyoti Mukherjee, Ryan A. Rossi, Anup Rao, Alexa Siu, and Sunav Choudhary.
Education
- 1. PhD in Artificial Intelligence (GSAI), KAIST, 2023-present, advised by Se-Young Yun (OSI Lab) and Chulhee “Charlie” Yun (OptiML Lab).
- 2. MSc in Artificial Intelligence (GSAI), KAIST, 2021-2023.
- 3. BSc in Mathematical Sciences and Computer Science (double major, cum laude), KAIST, 2017-2021.
Background
- Research interests include interactive machine learning and statistical analysis of large networks, with a focus on community detection. Broader interests encompass all aspects of mathematical and theoretical AI, as well as related problems in mathematics. Recently, also looking into various aspects of LLMs (alignment, reasoning, test-time scaling) from a theoretician's viewpoint.
Miscellany
- Enjoys organizing seminars and workshops to foster Korea’s theoretical AI community; an avid violinist who has performed as a first violinist with the KAIST Orchestra (KAO), PASSIONATE Orchestra, MDOP Orchestra, KAO-S, PROJECT Starlight, and PROJECT IDEA (expected).
