Kiljae Lee

Kiljae Lee

Ph.D. Candidate in Statistics

About Me

Hi! I am a Ph.D. candidate in Statistics at the Ohio State University, advised by Prof. Yuan Zhang.

I am passionate about advancing data-centric AI through principled methods for data attribution. Currently I am designing tools that quantify the value of dataset both fairly and efficiently [1]. My research focuses on developing theoretically grounded yet scalable approaches that can be applied across diverse AI applications including generative AI modeling.

As a statistician, I am also deeply interested in methods that quantify uncertainty in black-box models, especially those that remain robust under model misspecification [2]. My broader goal is to create algorithms that are not only statistically sound but also computationally feasible, so they can scale with modern AI systems.

Research Interests

My research interests span across several areas in statistics and machine learning:

  • Data-Centric AI
  • Uncertainty Quantification

Publications

  • Faithful Group Shapley Value [PDF]
    Kiljae Lee*, Ziqi Liu*, Weijing Tang and Yuan Zhang
    NeurIPS 2025
    ICML 2025 DataWorld Workshop, Best Paper Award (Honorable Mention; 3 out of 55 submissions)
    Developed a fast and robust method for group-data valuation that satisfies a novel faithfulness axiom, ensuring resistance to adversarial manipulation and strong theoretical guarantees.

  • Leave-One-Out Stable Conformal Prediction [PDF]
    Kiljae Lee and Yuan Zhang
    ICLR 2025
    Proposed a novel conformal prediction algorithm leveraging algorithmic stability, achieving fast and reliable uncertainty quantification without sacrificing statistical validity.

(* co-first author)

Experience

  • Statistical Consultant, The Ohio State University (Apr 2025 – Aug 2025)
    Advised 6 clients across psychology, special education, city planning, and agricultural science. Provided statistical modeling and analysis for dissertation defenses and grant proposals.

  • Graduate Teaching Assistant, The Ohio State University (Aug 2023 – Present)
    TA for 11 courses including Advanced Theory of Statistics, Statistical Computation, and Statistical Machine Learning.

  • Graduate Research Assistant, Korea University (Mar 2020 – Aug 2021)
    Developed a Bayesian semiparametric two-stage meta-analysis model and applied it to U.S. COVID-19 data (Master’s Thesis, manuscript in submission).

Education

  • The Ohio State University, Columbus, OH
    Ph.D. in Statistics (Aug 2022 – Expected Jun 2027)
    Advisor: Dr. Yuan Zhang

  • Korea University, Seoul, South Korea
    Master in Statistics (Mar 2020 – Aug 2021)
    Advisor: Dr. Taeryon Choi
    Thesis: Fully Bayesian Semiparametric Two-stage Meta-analysis

  • Korea University, Seoul, South Korea
    Bachelor of Statistics (Mar 2014 – Feb 2020)

Honors & Awards

  • Ransom & Marian Whitney Award for Research, Department of Statistics, The Ohio State University (2026); Recognizes independence, creativity, originality, progress, and potential for publication/application in PhD research. One of two awardees.
  • University Fellowship, The Ohio State University (2022–2023)
  • Academic Excellence Scholarships, Korea University (2015–2019)

Skills

Languages/Tools: Python (Scikit-learn, PyTorch, Keras, TensorFlow), R, LaTeX, SQL, C++, MATLAB, SAS