Akiko Takeda
Akiko Takeda completed a doctorate at Tokyo Institute of Technology in March 2001 and then worked in Toshiba’s research centre. She held appointments at Tokyo Tech, Keio, the University of Tokyo, and the Institute of Statistical Mathematics. She led RIKEN’s Continuous Optimization Team from September 2016 and became a professor at Tokyo in April 2018. Her research develops methods for nonconvex and uncertain optimisation and their use in machine learning. Her publications include work on accelerated gradient methods and hyperparameter selection, providing concrete examples of this mathematical research.
- Now at
- University of Tokyo and RIKEN
- Works
- 222
- Citations
- 1,717
- h-index
- 20
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
2001–2003
Doctoral research and Toshiba
Takeda earned a doctorate in mathematical and computing sciences in March 2001. She worked as a researcher at Toshiba’s corporate R&Dcentre from April 2001 to March 2003.
-
2008–2016
Academic optimisation research
She held assistant and associate professorships at Keio from 2008 to 2013 and an associate professorship at Tokyo from 2013 to 2016. Her research examines models and algorithms for difficult optimisation problems. Applications include machine learning, control, and decisions that must account for uncertainty.
-
2016–2018
RIKEN and Tokyo
She led RIKEN’s Continuous Optimization Team from September 2016. She became a professor at Tokyo in April 2018. The team studies efficient algorithms for nonconvex and uncertain problems, including the mathematical problems that arise when training learning systems.
-
2024
Gradient methods and hyperparameters
Takeda coauthored 2024 work on parameter-free accelerated gradient descent for nonconvex minimisation. She also coauthored work on hyperparameter selection using bilevel optimisation.
Sources and credits
Photo: www.riken.jp; 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.