Magnetic Architectures for Reservoir Computing Hardware (MARCH)

Research Project

Funders Involved

MARCH: Magnetic Architectures for Reservoir Computing Hardware

MARCH aims to investigate design and performance possibilities of in materio reservoir computing (RC), a form of unconventional computation that is particularly well-suited to handling complex time-varying data.

The project uses networks of interconnected magnetic nanorings, within which natural stochastic and dynamic processes combine to create the complex behaviours required for RC, e.g. nonlinear response and fading memory. The platform offers a flexible and scalable testbed for exploring and demonstrating novel experimental RC configurations and implementations.

The project also uses bespoke high performance software simulation and advanced search techniques to explore and evaluate the complex design space before committing to device fabrication.

The magnetic hardware platform and software design tools will be brought together to create a sophisticated demonstrator, using multiple reservoir components as sensors, processors and controllers in a mobile robotic platform.

MARCH brings together researchers from the Universities of York and Sheffield with expertise in Machine Learning, Reservoir Computing, Robotics, and Nanoscale Magnetism. This allows us to explore developing new multi-RC design tools, and will design, test and manufacture nanomagnetic RCs.

This will result in a new design methodology and platform for multi-reservoir devices, that can be exploited to design low-power, robust, flexible, and efficient smart sensing and other `edge-computing' devices in a diverse range of materials.