Professor: Kenji Doya

The Neural Computation Unit pursues the dual goals of developing robust and flexible learning algorithms and elucidating the brain’s mechanisms for robust and flexible learning.

Our specific foci are on reinforcement learning for a biological or artificial agent learns novel behaviors in uncertain environments by exploration and reward feedback, and dynamic Bayesian inference for grasping the state of theenvironment by combining model-based predictions and sensory evidence.

We combine top-down, theoretical approaches and bottom-up, experimental approaches to achieve these goals.

Research Projects

We collaborate with a variety of partners to run joint research, in addition to individual curiosity-driven research projects.

Optical Quantum Computing (2026-2031)

• Unified Theory of Prediction and Action (2023-2028)

• Brain/MINDS 2.0 Digital Brain Project (2024-2030)

• Fugaku Whole Brain Data Assimilation Project (2023-2026)

Global Bioconvergence Center of Innovation (2022-2026)

Online Textooks

Two textbooks based on courses at OIST are under development as Jupyter books with Python codes:

Research Positions

We are recruiting research staff for revealing how the brain works and creating brain-inspired adaptive systems.

We specifically invite applications to the following research topics:

For the following research topics, we currently do not have an opening but accept applicants with own funding:

If you are interested, please register your CV, publication list, research statement, and the names and e-mail addresses of up to three referees using the application form.

If you are interested to join as a PhD student or an intern, please apply from the Graduate School admission site