Research

Causal Inference

• Scalable Multi-Action Offline Policy Learning with an m-ary Tree
  KDD, 2026.
  A scalable offline policy learning method for multi-action m-ary policy trees that produces more expressive and interpretable policy trees than binary trees while scaling to industry-size datasets.

Natural Language Processing

• Keyword-Assisted Topic Models (with Kosuke Imai and Tomoya Sasaki)
  American Journal of Political Science, 2024.
  A keyword-assisted topic model that uses researcher-provided keywords to produce more interpretable, objective, and accurate topics while supporting covariates and time trends.
• Scale-Invariant Infinite Hierarchical Topic Model (with Daichi Mochihashi)
  Findings of ACL, 2023.
  A scale-invariant infinite hierarchical topic model that builds deeper, more diverse, and less redundant topic trees from corpora.

Applied Research

• Constitutional Design Preferences: An Experimental Approach (with Kenneth Mori McElwain, Ai Takahashi, and Christian G. Winkler)
  International Journal of Constitutional Law, 2024.
  A conjoint analysis showing that Japanese citizens prefer detailed constitutions, especially human rights provisions, and view amendment flexibility as a strategic safeguard.
• Just a Number? Voter Evaluations of Age in Candidate Choice Experiments (with Daniel M. Smith)
  The Journal of Politics, 2022
  A meta-analysis and conjoint experiment showing that voters across democracies consistently penalize older political candidates, even when age differences are substantively small.
• The Proposer or the Proposal? An Experimental Analysis of Constitutional Beliefs (with Kenneth Mori McElwain and Christian G. Winkler)
  Japanese Journal of Political Science, 2021
  A survey experiment on constitutional amendment preferences showing that Japanese citizens favor non-partisan proposals, while constitutional ideology shapes support across issue areas independently of partisanship.

Work in Progress

• Polyunsaturated Fatty Acids Promote Appetite via the Microbiome-Gut-Brain Axis (with Yi Jia Liow, Mustafa Talay, Vladimir Yeliseyev, Lynn Bry, and Rachel N. Carmody)
  bioRxiv preprint, 2025.
  A diet-microbiome-gut-brain model showing that specific industrialized diet components, especially PUFAs, can increase appetite by altering gut microbiota and brain serotonin metabolism.

Software

• keyATM an R package for Keyword-Assisted Topic Models.
• ihLDA: a Python implementation of the Scale-Invariant Infinite Hierarchical Topic Model.
• RcppJagger: an R package for Jagger, a fast, accurate, and space-efficient morphological analyzer.