Talks

Invited talks, tutorials, and research presentations. For lab reading groups and seminar materials, see Lab Seminars.

2026

  • Provably Efficient Regularized Online RLHF with Generalized Bilinear Preferences

    Invited Talk at Chung-Ang University, Seoul, Republic of Korea

    Host: Prof. Kyoungseok Jang

2025

  • A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits

    Top Conference Session (KSC 2025), Yeosu EXPO Convention Center, Yeosu, Republic of Korea

  • Bandits 101

    Invited Seminar at IIDS Lab, KAIST, Daejeon, Republic of Korea

  • A Primer on Bandits and Two Recent Theoretical Advances in Generalized Linear Bandits

    Invited Talk at Department of Statistics, Jeonbuk National University, Jeonju, Republic of Korea

    Host: Prof. Gwangsu Kim

  • Two Theoretical Advances on Sample-Efficient Preference Learning

    Reading Group at UCL, online

    Hosts: Prof. Ilija Bogunovic & Dr. Sangwoong Yoon

2024

  • Bandits 101 @Hyperconnect: A Maximally Non-technical Tutorial

    Invited Talk at Hyperconnect, Asem Tower, Seoul, Republic of Korea

    Host: Dr. Gihun Lee

  • A Unified Confidence Sequence for Generalized Linear Models, with Applications to Bandits

    Invited Session: Bandit and Reinforcement Learning Theory (CKAIA 2024), BEXCO, Busan, Republic of Korea

    Host: Prof. Min-hwan Oh

  • On Various Statistical Problems Arising from Preference Learning and RLHF + Misc. Theories

    SNU-KAIST AI/ML Theory Workshop, St. John's Hotel, Gangneung, Republic of Korea

  • Nearly Optimal Latent State Decoding in Block MDPs

    Learning in Networks: Structure, Dynamics, and Control (part of Digital Futures Focus Period Program), KTH, Stockholm, Sweden

    Host: Prof. Alexandre Proutière

  • Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion, and Beyond

    DeLTA Seminar, University of Copenhagen, Copenhagen, Denmark

    Host: Prof. Mohammad Sadegh Talebi

  • Nearly Optimal Latent State Decoding in Block MDPs

    Junior Researcher Session, Summer School on Optimal Transport, Stochastic Analysis, and Application to Machine Learning, Daejeon, Republic of Korea

  • Bandits 101 + Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion

    Invited Talk at ITML Lab, Yonsei University, Seoul, Republic of Korea

    Host: Prof. Jy-Yong Sohn

  • Improved Regret Bounds of (Multinomial) Logistic Bandits via Regret-to-Confidence-Set Conversion

    “Mini RL Theory Workshop”, Universitat Pompeu Fabra, Barcelona, Spain

    Host: Prof. Gergely Neu

2023

  • Fair Streaming Principal Component Analysis: Statistical and Algorithmic Viewpoint

    Top Conference Session (KSC 2023), Busan, Republic of Korea

  • Nearly Optimal Latent State Decoding in Block MDPs

    Invited Session on “Stochastic learning in structured systems”, 21st INFORMS Applied Probability Society (APS) Conference, Nancy, France

    Host: Prof. Jaron Sanders

  • Nearly Optimal Latent State Decoding in Block MDPs

    Top Conference Session (KCC 2023), Jeju Island, Republic of Korea

2022

  • Fast and Efficient MMD-based Fair PCA via Optimization over Stiefel Manifold

    Top Conference Session (KSC 2022), Jeju Island, Republic of Korea

  • Fast and Efficient MMD-based Fair PCA via Optimization over Stiefel Manifold

    KAIST Math Graduate Student Seminar (KMGS), online

  • Fast and Efficient MMD-based Fair PCA via Optimization over Stiefel Manifold

    KAIST Math, online

    Host: Prof. Donghwan Kim

2021