Tuning Physical Reservoir Computing for Edge Computing with Ionic Liquids

Ionic Liquid Reservoir Computers: Flexible and Efficient Edge Computing

Researchers in Japan have developed a physical reservoir that can be adjusted based on the dielectric relaxation of an electrode-ionic fluid interface.

In the near-future, artificial intelligence processing may need to be done at the edge – close to the user where data is collected and not on a remote computer server. It will be necessary to have high-speed processing of data with low energy consumption. This is why physical reservoir computing has become a popular platform. A recent breakthrough by scientists in Japan made it more flexible and useful.

Physical reservoir computing (PRC), a machine learning framework which uses the transient responses of physical systems to process time series signals quickly and efficiently, can be attractive because it is low-power. PRC systems are limited in their ability to be tuned, which limits the types of signals they can process. Researchers from Japan have developed ionic fluids that are easily tunable and can process signals on a wide range of timescales.


Ionic Liquid-Based Reservoir Computers: Efficient and Flexible Edge Computing