Machine learning researcher

I am a master's student in the School of Computing at KAIST, advised by Prof. Sungjin Ahn at MLML. I received my B.S. in Computer Science from KAIST.

My research interests lie in world models, theory learning, compositional generalization, and explanation-driven learning. I am especially interested in how models can form structured abstractions from raw observations and reuse them to explain new phenomena.

Recently, I have been working on Learning to Theorize, a framework for inferring executable theories from observations without direct supervision of the underlying causes.

News

  • May 2026
    Our paper Learning to Theorize the World from Observation was selected as an ICML 2026 Spotlight.
  • Mar 2026
    Our paper Extendable Planning via Multiscale Diffusion was selected for an oral presentation at AAAI 2026.
  • Jul 2025
    Our paper Monte Carlo Tree Diffusion for System 2 Planning was selected as an ICML 2025 Spotlight.

Selected Publications

2026

  • ICML

    Learning to Theorize the World from Observation

    Doojin Baek*, Gyubin Lee*, Junyeob Baek, Hosung Lee, Sungjin Ahn

    International Conference on Machine Learning, 2026. Spotlight

    Project / arXiv / PDF

  • AAAI

    Extendable Planning via Multiscale Diffusion

    Chang Chen*, Hany Hamed*, Doojin Baek, Taegu Kang, Samyeul Noh, Yoshua Bengio, Sungjin Ahn

    AAAI Conference on Artificial Intelligence, 2026. Oral Presentation

    arXiv

2025

  • ICML

    Monte Carlo Tree Diffusion for System 2 Planning

    Jaesik Yoon, Hyeonseo Cho, Doojin Baek, Yoshua Bengio, Sungjin Ahn

    International Conference on Machine Learning, 2025. Spotlight

    arXiv

2024

  • ICML

    Enforcing Constraints in RNA Secondary Structure Predictions: A Post-Processing Framework Based on the Assignment Problem

    Geewon Suh, Gyeongjo Hwang, Seokjun Kang, Doojin Baek, Mingeun Kang

    International Conference on Machine Learning, 2024

    Paper / OpenReview