Shinji Ito
Shinji Ito studies sequential decisions, online learning and bandit algorithms. These methods choose actions while receiving incomplete feedback about other available choices. He directs RIKEN’s Sequential Decision Making Team, and his shared research covers feedback graphs, partial monitoring and robust algorithms. His 2023 paper with Kei Takemura develops a linear-bandit algorithm that adapts to different kinds of environment and to properties such as variation and noise. His work concerns theoretical guarantees under stated assumptions.
Timeline
-
2022
Feedback and robust learning
Ito coauthored research on online learning with feedback graphs and a paper on online submodular minimization. These examine how a learner makes decisions with restricted observations and changing losses.
-
2023
Adaptive linear-bandit algorithm
Ito and Kei Takemura coauthored a COLT paper on a linear-bandit algorithm with variance-adaptive regret bounds. The original abstract describes adaptation to adversarial environments and stochastic settings with corruption. Regret measures accumulated loss relative to a comparison action.
-
2023
Other forms of limited feedback
Ito’s work also includes partial monitoring, combinatorial semi-bandits and task assignment with unknown processing time. These are distinct decision problems that require learning from limited information.
-
As of October 2026
Sequential Decision Making Team
Ito directs the team. The team studies online learning, bandit problems and reinforcement learning for uncertain and fluctuating environments. The programme seeks theory and algorithms.
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.