About Me

Welcome to my website. I am a Ph.D. candidate at Harvard Kennedy School (Politics and Institutions track).

My research centers on political methodologies and applications. Methodologically, I am interested in the intersection of causal inference and machine learning, especially causal inference methods for texts, images, and videos. Substantively, my focus lies on political communication and misinformation. See my research page for the current research program.

I have contributed to several research projects aligned with these interests, and my work has been published in peer-reviewed journals, including Journal of the American Statistical Association, Proceedings of the National Academy of Sciences, Journal of Conflict Resolution, and the International Journal of Communication.

Before joining Harvard, I earned an M.S. in Statistics from the University of Chicago and a B.A. in Political Science from Waseda University in Japan. Feel free to reach out to me at knakamura [at] g.harvard.edu.

Recent Updates

• 2026/08/04: GenAI-Powered Inference is accepted at Proceedings of the National Academy of Sciences
• 2026/08/03: Presented A General Approach to Correcting Measurement Errors When Only Indirect Validation is Possible (with Naoki Egami) at Joint Statistical Meeting
• 2026/07/19: Presented the followings at Society for Political Methodology Annual Meeting.
  • Causal Inference with Video Features as Treatments (with Adam Breuer, Michael H. Crespin, Bryce J. Dietrich, Kosuke Imai)
  • Where’s the Evidence that Respondents Understand Your Survey Questions? (with Musashi Hinck, Gary King, and Brandon Stewart)
  • Identification and Estimation for AI- and Algorithm-Powered Interventions (Poster)
  • No “Human Labels as Gold Standard”: When LLMs Disagree, Humans Do Too (Poster, with Jing Ling Tan and George Yean)
• 2026/07/08: Released a new working paper Causal Inference with Video Features as Treatments with Adam Breuer, Michael H. Crespin, Bryce J. Dietrich, Kosuke Imai
• 2026/07/01: Released a new working paper Where’s the Evidence that Respondents Understand Your Survey Questions? with Musashi Hinck, Gary King, and Brandon Stewart
• 2026/06/01: Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments is accepted at Journal of the American Statistical Association
• 2026/05/10: Released a new working paper GenAI Powered Dynamic Causal Inference with Unstructured Data with Kosuke Imai
• 2026/04/03: Invited Talk: Frontier in Political Methodology (at Washington University at St. Louis)
• 2026/02/23: Invited Talk: Stanford University (Guest speaker for POLISCI 450D)
• 2026/02/10: Invited Talk: Causal Inference Seminar, Harvard University
• 2025/09/17: Presented Surrogate Representation Inference for Noisy Text and Image Annotations at Applied Statistics Workshop, Harvard University
• 2025/09/15: Released a new working paper Surrogate Representation Inference for Noisy Text and Image Annotations
• 2025/09/13: Discussant: Annual Meeting of American Political Science Association
• 2025/09/12: Presented GenAI-Powered Inference at Annual Meeting of American Political Science Association
• 2025/07/25: Received the best poster award (method) at Society for Political Methodology Annual Meeting.
• 2025/07/18: Presented GenAI-Powered Inference (with Kosuke Imai, Presentation) and Efficient Inference for Text-as-Outcome from Partial and Noisy Annotations (Poster) at Society for Political Methodology Annual Meeting.
• 2025/07/07: Released a new working paper GenAI-Powered Inference with Kosuke Imai
• 2025/06/28: Invited Talk: Nospare
• 2025/05/20: Presented a tutorial of gpi_pack at Oxford Computational Political Science Group (slide is available from here)
• 2025/02/27: Released the software gpi_pack for generative-AI powered statistical inference
• 2025/02/12: Presented Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments at Applied Statistics Workshop, Harvard University
• 2025/01/05: Presented Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments at JSQPS Winter Meeting
• 2024/10/03: Released a new working paper Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments with Kosuke Imai
• 2024/08/01: Understanding the Impact of Military Service on Support for Insurrection in the United States is published on Journal of Conflict Resolution
• 2024/07/19: Presented Automated Cognitive Debriefing (with Musashi Hinck, Uma Ilavarasan, Gary King, and Brandon M. Stewart) and Causal Inference with Unstructured High-Dimensional Treatments using Deep Generative Models: Applications to Images and Texts (with Kosuke Imai) at Society for Political Methodology Annual Meeting.