📧E-Mail confeitohs gmail hslee kaist 📞Phone 010-8562-0096

**GitHub

CV

Google Scholar**

😊Hi,

I am a first-year Master’s student at KAIST School of Computing. I am currently in MLML @ KAIST led by Professor Sungjin Ahn. My current research/study interests are:

• Artificial General Intelligence
  • Bridging Perception ↔ Cognition
  • System-2 Reasoning
  • Compositional Representation
• Learning Environments & Benchmarks
  • ARC-AGI

Additionally, I am used to utilizing several skill stacks to implement what I want. Here are the skills that I have used.

• Python, C/C++, Typescript + C#, HTML5
• PyTorch for Machine Learning
• Node.js - ExpressJS, NestJS for Web Backend
• Premiere Pro and After Effects (Basic)

📑Publication

Discrete JEPA: Learning Discrete Token Representations without Reconstruction

Junyeob Baek, Hosung Lee, Chris Hoang, Mengye Ren, Sungjin Ahn ICML 2025 Tokenization Workshop (TokShop). arXiv

Addressing and Visualizing Misalignments in Human Task-Solving Trajectories

Sejin Kim, Hosung Lee, Sundong Kim KDD 2025. arXiv

ARCLE: The Abstraction and Reasoning Corpus Learning Environment for Reinforcement Learning

Hosung Lee*, Sejin Kim*, Seungpil Lee, Sanha Hwang, Jihwan Lee, Byung-Jun Lee, and Sundong Kim CoLLAs 2024. arXiv GitHub (Extended version of ARCLE paper in KSC)

O2ARC 3.0: A Platform for Solving and Creating ARC Tasks