Neuromorphic Devices Group
Semiconductor Materials Field
We develop high-performance neuromorphic devices that leverage the spatiotemporal dynamics of materials (ions, electrons, spins).
Group Leader:Takashi Tsuchiya
Two Orders of Magnitude Reduction in Computational Load Achieved by Ultrawideband Responses of an Ion-Gating Reservoir
Current Topics
We developed a physical reservoir device, ion-gating reservoir (IGR), utilizing ultrawide and complex responses of ion and electron at the graphene/ion-gel interface, achieving high computational performance comparable to deep learning while reducing the computational load by 99%. The IGR offers low-power, high-performance computing for resource-constrained environments, such as edge AI.
A schematic diagram of the ion-gating reservoir (IGR) composed of an ion-gel and monolayer graphene (left) and relationship between computational load and prediction error for the 10-step-ahead prediction task for the Mackey-Glass equation (right).
References
Group members
-
Takashi Tsuchiya
- Group Leader
-
Yoshitaka Shingaya
- Senior Researcher
-
Ryo Iguchi
- Senior Researcher