이호준 (Hojun Yi, 172635)

Hello, I am a fourth-year undergraduate student majoring in Computer Science and Mathematical Sciences at the Gwangju Institute of Science and Technology (GIST). I am currently working as an undergraduated researcher in the DataScience Lab, advised by Prof. Sundong Kim.

My ultimate research goal is to realize robust Artificial General Intelligence (AGI) capable of autonomously analyzing underlying rules and adapting to unseen Out-of-Distribution (OOD) environments by leveraging prior knowledge and preference/dispreference awareness.

To achieve this, my current research interests include below two core directions:

1. Autonomous Exploration & Continual Learning in RL

To maximize agent autonomy and learning efficiency, focusing on how agents can adapt smoothly with minimal data.

• Unsupervised Environment Design (UED): Developing highly efficient agent architectures that autonomously generate curriculum frameworks based on environmental difficulty and data distribution.
• Mitigating Catastrophic Forgetting: Utilizing Optimal Transport (OT) theory and Flow Policies to enhance policy and data transfer efficiency across diverse environmental distributions, enabling sustainable knowledge expansion.

2. Geometric Foundations of Latent Spaces & Explainable AI

To clarify the semantic relationships and complex rules embedded within high-dimensional data, connect linguistic/semantic consensus with the geometric structures of latent spaces.

• Statistical Manifolds & Information Geometry: Moving beyond Euclidean spaces, model latent representations as statistical manifolds. By applying topology and information geometry, mathematically define the curvature and metrics of probability distributions.
• Mathematical Causality & Interpretability: By constructing abstract semantics based on geometric structures, provide interpretability and causality for decision-making systems.

Ultimately, I aim to break through the current limitations of AI by leveraging on mathematical logic.

I am always open to intellectual discussions and exchanging ideas-feel free to reach out!

Education

• B.S. in Electrical Engineering and Computer Science GIST(Gwangju Institute of Science and Technology) 2023 - Present
  • Expected Cum Laude graduation in Aug 2027, Total GPA: 4.20/4.5
  • Major(EECS) GPA: 4.34, Double Major(Mathematical Sciences): 4.29

Experience

• Undergraduated Research Intern - DataScience Lab (Advised by Prof. Sundong Kim) in GIST 2025 - Present
  • Conducting Research on Open-Endedness, Unsupervised RL
  • Studied Math for ML, Probabilistic ML
  • Lightning talk about TRACED on CoRL workshop 2025

Selected Publications

All Publications

• TRACED: Transition-aware Regret Approximation with Co-learnability for Environment Design
  • Authors : G. Cho, J. Im, J. Lee, H. Yi, S. Kim, S. Kim
  • ICLR 2026, CoRL workshop 2025
  • [Paper] [Code] [Website]

Awards and honors:

• Dean’s List (GIST EECS) Spring 2025
• 2024 ICPC Seoul Regional - 63rd 2024-11
• 2024 GIST Algorithm Masters - 1st 2024-05
• Hello, BOJ 2024! - 45th 2024-01
• 2023 ICPC Seoul Regional - 49th 2023-11
• Academic Excellence Scholarship (GIST) 2023 - 2025
• Government Sponsored Scholarship (GIST) 2023 - Present