Taehui Yun

I am a graduate student in the Causality Lab at Seoul National University , advised by Prof. Sanghack Lee. I received my B.S. in Industrial and Management Engineering from POSTECH .

My research goal is to build convincing and safe AI systems grounded in causality. I am particularly interested in causal generative models, causal explainability (XAI), and time-series. Recently, I have been studying causal normalizing flows for time-series data and concept based model with causal lens. I enjoy deep-diving into diverse causal topics and collaborations. Please feel free to contact me if you share similar interests!

Email  /  GitHub  /  LinkedIn

Education

Experience

Research

  • Research Intern, Causality Lab, Seoul National University (2024.09 – 2025.02)
    Study on causal normalizing flows for time-series data
    Advisor: Prof. Sanghack Lee

  • Research Intern, FIRM Lab, POSTECH (2022.05 – 2022.09)
    Study on hierarchical risk parity
    Advisor: Prof. Bong-Gyu Jang

  • Research Intern, EDEN Lab, POSTECH (2020.12 – 2021.09)
    UX design for blockchain systems
    Advisor: Prof. Sung H. Han

Industry

  • Engineer Intern, LG Electronics, Seoul, Korea (2024.06 – 2024.09)
    Development of an open-source LLM utilization library and PEFT modules

  • Engineer Intern, GNE, Seoul, Korea (2023.03 – 2023.09)
    Development of a financial risk management platform

  • Engineer Intern, JURO Instruments, Seoul, Korea (2023.01 – 2023.03)
    Database analysis using financial engineering methods

Awards & Honors

  • Mirae Asset International Exchange Scholarship, Mirae Asset Group (2024)
  • IME Global Scholarship, POSTECH (2023)
  • Exchange Program Scholarship, POSTECH (2023)
  • CES 2023 Conference Scholarship, POSTECH (2023)
  • Student Advisor Scholarship, POSTECH (2022)
  • Jigok Scholarship, POSTECH (2020 – 2025)

Projects

  • Scalable Causal Discovery (2025.09 – Present)
    LG AI Research, Seoul, Republic of Korea

  • Causal Deep Generative Models for Time-Series (2024.09 – 2025.09)
    LG AI Research, Seoul, Republic of Korea

Publications

* indicates joint first authorship (equal contribution), † indicates corresponding author.

  • Breaking Bad: Component-Wise Parent Deletion for Score-Based Causal Discovery
    Min Woo Park*, Taehui Yun*, YoungIn Jang, Yoonseok Yeom, Jonghwan Kim, Jiyeon Kang, Songseong Kim, Hyemin Jung, Sangmin Lee, Jongseong Jang†, Sanghack Lee†
    Under review

  • Estimating Interventional Outcomes over Time with Causal Normalizing Flow
    Yoonseok Yeom*, Jonghwan Kim*, Taehui Yun, Juhyun Lyu, Jung-Hee Kim, Sangmin Lee, Jinseok Yang, Hyemin Jung, Woohyung Lim†, Sanghack Lee†
    Under review

This website is based on the template from here. here.