Minjae Oh

About

I am a Ph.D. student in Data Science at Seoul National University, advised by Prof. Yohan Jo in the Human-Oriented Language Intelligence (HOLI) Lab. My research centers on reinforcement learning and reasoning with large language models. Recently, I have been looking into physical AI and world models.

Education

• Seoul National University
  Ph.D. in Data Science (Advisor: Prof. Yohan Jo) 2026 Mar. – Present
• Seoul National University
  M.S. in Data Science (Advisor: Prof. Yohan Jo) 2024 Mar. – 2026 Feb.
• Korea University
  B.S. in Industrial Management Engineering 2018 Mar. – 2024 Feb.

1. The Low-Rank Structure of VLA Reinforcement Learning
   Minjae Oh^*, Yoonah Park^*, Jongwon Lim^*, Yohan Jo
   Under Review, 2026 paper
2. KL for a KL: On-Policy Distillation with Control Variate Baseline
   Minjae Oh, Sangjun Song, Gyubin Choi, Yunho Choi, Yohan Jo
   Neural Information Processing Systems (NeurIPS), 2026 paper blog code
3. Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States
   Yunho Choi^*, Jongwon Lim^*, Woojin Ahn, Minjae Oh, Jeonghoon Shim, Yohan Jo
   Neural Information Processing Systems (NeurIPS), 2026 paper
4. SHAPE of Chain-of-Thought in Math Reasoning
   Jonghyun Song, Sangjun Song, Minjae Oh, Haesung Pyun, Sungsik Lee, Yohan Jo
   AI4MATH@ICML, 2026 paper
5. Future Policy Approximation for Offline Reinforcement Learning in LLM Reasoning
   Minjae Oh, Yunho Choi, Dongmin Choi, Yohan Jo
   Under Review, 2026 paper
6. ThinkBrake: Efficient Reasoning via Log-Probability Margin Guided Decoding
   Sangjun Song^*, Minjae Oh^*, Seungkyu Lee, Sungmin Jo, Yohan Jo
   Findings of the Association for Computational Linguistics (ACL Findings), 2026 paper code
7. Medication Recommendation for Parkinson's Disease Based on Dynamics of Symptom Progression
   Minjae Oh, Hongbum Kim, Hyo Kyung Lee
   Scientific Reports, Vol. 14, Article 25051, 2024 paper