Ichiro Takeuchi
Ichiro Takeuchi develops machine learning methods that help scientists test discoveries and judge whether patterns in data have statistical support. Takeuchi coauthored work on measuring the statistical significance of image segmentation produced by a neural network. Other shared research in the same year addressed selective inference for generalized lasso models and methods for computing conformal prediction sets. RIKEN names Takeuchi as director of its Data-Driven Experimental Design Team. The team develops learning methods for scientific research and works to make data-driven experiments more reliable, transparent and reproducible.
Timeline
-
2022
Testing learned image patterns
Takeuchi coauthored work on measuring the statistical significance of image segmentation produced by a neural network. Other shared research in the same year addressed selective inference for generalized lasso models and methods for computing conformal prediction sets.
-
2023
Using data in scientific studies
Takeuchi coauthored work on statistically significant patterns in allergen protein sequences, image retrieval for lymphoma tissue and sparse rule models. These studies connected statistical methods with biological and medical data.
-
2024
Examining attention and adaptation
Takeuchi coauthored a statistical test for attention maps in Vision Transformers. Another paper addressed anomaly detection after domain adaptation. His work shows a continuing concern with checking patterns produced by machine learning.
-
As of October 2026
Data-driven experimental design
RIKEN names Takeuchi as director of its Data-Driven Experimental Design Team. The team develops learning methods for scientific research and works to make data-driven experiments more reliable, transparent and reproducible.
Sources and credits
Photo: www.riken.jp; Ichiro Takeuchi profile photograph; creator not stated. License: Not stated by source. Source: riken.jp. Cropped to a square and resized. Ask us to remove or replace this photo.