Scholar
Jeonghye Kim
Google Scholar ID: koDFScAAAAAJ
PhD candidate, KAIST
Citations & Impact
Citations
49
H-index
4
i10-index
1
Publications
10
Co-authors
18
list available
Contact
Publications
Resume (English only)
Academic Achievements
- - Publications:
- - [U1] Exploratory Memory-Augmented LLM Agent via Hybrid On- and Off-Policy Optimization
- - [C8] RL-Studio: A System for Multi-Phase Reinforcement Learning Experimentation
- - [W2] RAISE: Enhancing Scientific Reasoning in LLMs via Step-by-Step Retrieval
- - [C7] ReflAct: World-Grounded Decision Making in LLM Agents via Goal-State Reflection
- - [W1] Align While Search: Belief-Guided Exploratory Inference for Test-Time World Alignment
- - [C6] Penalizing Infeasible Actions and Reward Scaling in Reinforcement Learning with Offline Data
- - [C5] Online Pre-Training for Offline-to-Online Reinforcement Learning
- - [C4] ARS: Adaptive Reward Scaling for Multi-Task Reinforcement Learning
- - [C3] Adaptive Q-Aid for Conditional Supervised Learning in Offline Reinforcement Learning
- - [C2] Decision ConvFormer: Local Filtering in MetaFormer is Sufficient for Decision Making
- - [C1] LESSON: Learning to Integrate Exploration Strategies for Reinforcement Learning via an Option Framework
- - Awards:
- - ICML 2025 Spotlight (Top 2.6%)
- - ICLR 2024 Spotlight (Top 5%)
Research Experience
- - Research Intern, Microsoft Research, Shanghai, China (2025.09-2026.02), Mentor: Xufang Luo
- - Research Intern, Machine Intelligence Lab, SNU, Seoul, South Korea (2025.03-2025.09)
- - Research Intern, LG AI Research, Seoul, South Korea (2024.03-2024.12), Mentor: Kanghoon Lee
- - CEO, Dearplants, Daejeon, South Korea (2020.06-2021.11)
Education
- - Ph.D. Candidate in Electrical Engineering, KAIST, Advisor: Prof. Youngchul Sung (2024.03~)
- - M.S. in Electrical Engineering, KAIST, Advisor: Prof. Youngchul Sung (2022.03-2024.02)
- - B.S. in School of Computing, KAIST (Cum laude) (2015.03-2022.02)
- - B.A. in Psychology, Bachelor's Degree Examination for Self-Education (BDES) (2018)
Background
- Research Interests: How reinforcement learning can enhance the reasoning and decision-making of intelligent agents, especially by improving pretrained policies such as large language models and vision-language models.
Miscellany
- Invited Talks:
- - 2025.02 "Adaptive Q-Aid for Conditional Supervised Learning in Offline Reinforcement Learning" @ ML2
- - 2023.10 "LESSON: Learning to Integrate Exploration Strategies for Reinforcement Learning via an Option Framework" @ CARAI
