Woocheol Shin

M.S. Student, Graduate School of Data Science, KAIST

prof_pic_new.jpg

I am an M.S. student in the Graduate School of Data Science at KAIST, advised by Prof. Jinkyoo Park in the System Intelligence Lab.

I received my B.S. in Computer Science and Engineering from UNIST. I worked with Prof. Junghoon Kim at the Data Mining Lab, focusing on graph-structured problems, particularly algorithm design and problem formulation.

I am broadly interested in building generative policies for sequential decision-making toward human-desired goals in the real world. My current research focuses on steering large generative models, especially diffusion and flow-based models, toward desired outputs using reinforcement learning and reward-guided optimization.

news

Jan 2026 Two papers accepted to ICLR 2026!
Oct 2025 Paper accepted to WSDM 2026!
Jun 2025 Paper accepted to Knowledge-Based Systems!
Feb 2025 Started M.S. at the Graduate School of Data Science, KAIST, advised by Prof. Jinkyoo Park.

selected publications

  1. Diffusion Fine-Tuning via Reparameterized Policy Gradient of the Soft Q-Function
    Hyeongyu Kang*, Jaewoo Lee*, Woocheol Shin*, and 2 more authors
    In International Conference on Learning Representations (ICLR), 2026
    *Equal contribution.
  2. Diffusion Alignment as Variational Expectation-Maximization
    Jaewoo Lee, Minsu Kim, Sanghyeok Choi, and 7 more authors
    In International Conference on Learning Representations (ICLR), 2026
  3. Urban Traffic Network Layout Optimization with Guided Discrete Diffusion Models
    Taeyoung Yun, Inhyuck Song, Woocheol Shin, and 4 more authors
    In ACM International Conference on Web Search and Data Mining (WSDM), 2026