Aarhus University Seal

SpinGrain

Spintronic nano-grains with local vortex oscillations for reservoir computing

Status: Start soon ...

  • start date

1 March 2022  

  • end date

28  February 2024 

Funded under: Villum Experiment 

  • Overall budget:

DKK  2M 

  •   Grant agreement ID: 

 TBD  

Coordinated by: AARHUS UNIVERSITET

Project Description

Artificial intelligence (AI) has been witnessed explosive growth over the past two decades. The core processing unit of AI is artificial neural networks (ANNs) inspired by the human brain with remarkable accuracy progress in cognitive applications; surpassed human brain in 2015 for the first time [1]. However, the biggest challenge of AI is the low performance of their hardware implementation using typical computers (CPUs/GPUs). The human brain consumes ~20W to perform more than 1016 operations, 1012 times better than the current supercomputer technology [2]. The most promising solution to reduce this huge gap is brain-inspired computing hardware.

In this project, I believe that I can take brain-inspired computing (BIC) systems a big step forward by scaling down the AI into a single nano-device for the first time. If it succeeds, we will have a BIC system in the size of a dust particle with energy consumption so small that can be harvested from its environment, open a new gate towards many new AI applications that were impossible in the past due to low performance of AI hardware such as spacecraft, edge computing, data telemetry in body, etc.