Profile Image

Hi! I am a PhD student in the Department of Statistics and Data Science at Carnegie Mellon University (CMU), where I am advised by Larry Wasserman and Sivaraman Balakrishnan. You can contact me at lucaskania@cmu.edu.

My research in statistical machine learning and non-parametric statistics aims to develop practical algorithms with provable guarantees. My current research interests include:

  • Non-asymptotic hypothesis testing: In high-dimensional settings, estimation may be infeasible, but we can still design algorithms that make reliable decisions.
  • Augmenting algorithms with predictions: Classical tests often fail to detect meaningful patterns in high dimensions. Incorporating machine learning predictions can overcome this limitation.
  • Invariant characterizations of causality: Causal frameworks link observed and unobserved data distributions. They can be exploited to design algorithms that perform well in unseen scenarios.

Currently, I am reviewing for the Journal of the American Statistical Association and Information and Inference.

My papers are available below and at Google Scholar.

News

I will present methods to optimally assess agreement across studies in meta-analyses at JSM 2025. Contact me if you plan to attend.

Preprints

(*) denotes equal contribution, and (Α-Ω) denotes alphabetical ordering.

Testing Random Effects for Binomial Data (Paper)
Lucas Kania, Larry Wasserman, and Sivaraman Balakrishnan

Robust semi-parametric signal detection in particle physics with classifiers decorrelated via optimal transport (Paper)
Purvasha Chakravarti*, Lucas Kania*, Olaf Behnke, Mikael Kuusela, and Larry Wasserman

Causal Regularization (Paper)
Lucas Kania and Ernst Wit

Conference Publications

Sparse Information Filter for Fast Gaussian Process Regression (Paper)
Lucas Kania, Manuel Schürch, Dario Azzimonti, and Alessio Benavoli
Machine Learning and Knowledge Discovery in Databases. Research Track. ECML PKDD 2021

Novel Range Functions via Taylor Expansions and Recursive Lagrange Interpolation with Application to Real Root Isolation (Paper)
(Α-Ω) Kai Hormann, Lucas Kania, and Chee Yap
Proceedings of the 2021 International Symposium on Symbolic and Algebraic Computation. ISSAC 2021