Statistical Science and Causal Discovery

I work on developing novel statistical methods to analyze data observed in social sciences, biomedical sciences, and natural sciences. The key is causal reasoning: Brief introduction and current research focus.

 Recent Highlights:

• NEW Non-Gaussian methods for causal structure learning (Shimizu, Prevention Science, 2018)
• Learning instrumental variables with structural and non-Gaussianity assumptions (Silva & Shimizu, JMLR, 2017)
• Estimation of interventional effects of features on prediction (Blöbaum & Shimizu, MLSP2017)
• Introductory slides
• Special feature on recent developments in causal discovery and inference (Winter, 2017)

 The LiNGAM Project:

This project develops semiparametric methods for learning mathematical models known as structural equation models. An important application of those methods is causal discovery.

• A short introduction
• NEW A tutorial paper (Prevention Science, 2018)
• Introductory book: Causal Discovery (2017, Kodansha, in Japanese) 
• 2018 Meiko Sugiyama Award (Publication Award) from the Behaviormetric Society
• Introductory book chapters (Chapter 7 and Chapter 8. In Statistics and Causality, 2016)
• A survey paper (Behaviormetrika, 2014)
• Tutorials:
  • Tutorial slides at BSJ2012 and references (In Japanese)
  • Tutorial slides at UAI2010 and references
  • Talk slides at IMPS2008 and erratum
• A collection of LiNGAM-related papers