PeopleofSuperintelligence

Researcher

Hidetoshi Shimodaira

Also known as 下平英寿

Hidetoshi Shimodaira is a statistician whose 2000 paper on predictive inference under covariate shift coined that term, and whose lab at Kyoto University now studies language models. He earned a doctorate in engineering at the University of Tokyo in 1995 and went on to publish phylogenetic tree tests and the CONSEL and pvclust software. After posts at the Institute of Statistical Mathematics, Tokyo Institute of Technology and Osaka University, he was a team leader at RIKEN's Center for Advanced Intelligence Project from 2016 to 2022 and became a Kyoto University professor in 2017. In 2025, his paper Mapping 1,000+ Language Models via the Log-Likelihood Vector won an Outstanding Paper award at ACL 2025. He is a professor in Kyoto University's Graduate School of Informatics and unit head of the university's Artificial Intelligence Research Unit.

Now at
RIKEN and Center for the Promotion of Interdisciplinary Education and Research
Badges

Timeline

  1. 1990–1995

    Degrees at the University of Tokyo

    He earns a bachelor's degree in 1990, a master's in 1992 and a doctorate in 1995, all in engineering from the University of Tokyo's Department of Mathematical Engineering and Information Physics.

  2. 1995–2002

    Institute of Statistical Mathematics

    He spends 1995 to 1996 as a postdoc in the Felsenstein laboratory at the University of Washington's Department of Genetics. From 1996 to 2002, he is an assistant professor at the Institute of Statistical Mathematics, and from 1999 to 2001 he is also a visiting scholar at Stanford's Department of Statistics with Brad Efron.

  3. 1999–2006

    Phylogenetic tests and pvclust

    In 1999, he and Hasegawa publish Multiple comparisons of log-likelihoods with applications to phylogenetic inference in Molecular Biology and Evolution, and their CONSEL software for assessing the confidence of phylogenetic tree selection follows in Bioinformatics in 2001. In 2002, he publishes An approximately unbiased test of phylogenetic tree selection in Systematic Biology. In 2006, he and Ryota Suzuki publish Pvclust, an R package that calculates p-values for hierarchical clusters through multiscale bootstrap resampling.

  4. 2000

    Coining covariate shift

    In 2000, he publishes Improving predictive inference under covariate shift by weighting the log-likelihood function in the Journal of Statistical Planning and Inference, the paper in which he coins the term covariate shift. In 2008, he contributes a chapter to the MIT Press volume Dataset Shift in Machine Learning.

  5. 2002–2012

    Associate professor at Tokyo Tech

    In 2002, he becomes a senior lecturer in the Department of Mathematical and Computing Sciences at Tokyo Institute of Technology, and he is an associate professor there from 2005 to 2012. He receives the Young Scientist Award from Japan's Ministry of Education, Culture, Sports, Science and Technology in 2006 and the Japanese Statistical Society Award for Research in 2010.

  6. 2012–2022

    Professor and RIKEN AIP team leader

    He is a professor in Osaka University's Graduate School of Engineering Science from April 2012 to March 2017. In September 2016, he becomes a team leader at RIKEN's Center for Advanced Intelligence Project, where his lab is the Mathematical Statistics Team until March 2022. In April 2017, he joins Kyoto University's Graduate School of Informatics as a professor.

  7. 2025

    An ACL Outstanding Paper award

    In 2025, Mapping 1,000+ Language Models via the Log-Likelihood Vector, written with Momose Oyama, Hiroaki Yamagiwa and Yusuke Takase, receives an Outstanding Paper award at ACL 2025. The paper uses log-likelihood vectors on a predefined text set as model features, so that squared Euclidean distance approximates the Kullback-Leibler divergence of text-generation probabilities, and applies the method to over 1,000 language models to build a model map.

  8. As of October 2026

    Professor and AI research unit head

    As of October 2026, he is a professor in Kyoto University's Graduate School of Informatics and unit head of the Artificial Intelligence Research Unit at the Center for the Promotion of Interdisciplinary Education and Research. He is also a visiting scientist at RIKEN's Center for Advanced Intelligence Project, a post he takes up in April 2022. A Grant-in-Aid for Scientific Research (A) project on log-likelihood vectors for model evaluation in the generative AI era runs from April 2026 to March 2030.

Career

  1. Since 2017Kyoto UniversityProfessor
  2. Since 2016RIKENTeam Leader, Visiting Scientist
  3. NowCenter for the Promotion of Interdisciplinary Education and ResearchUnit Head
  4. 2012–2017Osaka UniversityProfessor
  5. 2002–2012Institute of Science TokyoSenior Lecturer, Associate Professor
  6. 1999–2001Stanford UniversityVisiting Scholar
  7. 1996–2002Research Organization of Information and SystemsAssistant Professor
  8. 1995–1996University of WashingtonPostdoc

Education

  1. 1995University of TokyoDoctorate
  2. 1992University of TokyoMaster's degree
  3. 1990University of TokyoBachelor's degree

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Hidetoshi Shimodaira. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. Universities keep AI researchers for a median of 6 years, compared with 2 years at frontier labsIncludes RIKEN. Across 15,811 dated work stays (6,529 people), median role tenure runs from 6 years at universities to 2 years at frontier labs. We compare all five organization types, starting decades from the 1970s to the 2020s, and the organizations that keep people longest.
  3. The United States employs an estimated 48% of people shaping AIIncludes RIKEN. We estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.
  4. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes RIKEN. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Hidetoshi Shimodaira yet. Send us a free-licensed photo.

  1. Curriculum Vitae stat.sys.i.kyoto-u.ac.jp. Checked 21 October 2026.
  2. Shimodaira, Hidetoshi (Graduate School of Informatics, Department of Informatics) | Activity Database on Education and Research, Kyoto University kdb.iimc.kyoto-u.ac.jp. Checked 21 October 2026.
  3. Hidetoshi Shimodaira – Shimodaira Lab stat.sys.i.kyoto-u.ac.jp. Checked 21 October 2026.
  4. pvclust: An R package for hierarchical clustering with p-values github.com. Checked 21 October 2026.
  5. Kyoto University and RIKEN – Shimodaira Lab stat.sys.i.kyoto-u.ac.jp. Checked 21 October 2026.
  6. Mapping 1,000+ Language Models via the Log-Likelihood Vector - ACL Anthology aclanthology.org. Checked 21 October 2026.
  7. Awards - ACL 2025 2025.aclweb.org. Checked 21 October 2026.
  8. Outstanding Paper Award at the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025) – Shimodaira Lab stat.sys.i.kyoto-u.ac.jp. Checked 21 October 2026.