A New Paradigm of Reservoir Computing Exploiting Hydrodynamics

TL;DR

This paper introduces Aqua-PACMANN, a reservoir computing system using hydrodynamic waves in shallow water to perform complex tasks with low energy, demonstrated by an XNOR logic gate as proof of concept.

Contribution

It presents a novel fluid dynamic neuromorphic computing paradigm leveraging hydrodynamic waves, reducing electronic components to a simple detection camera.

Findings

Successful implementation of an XNOR logic gate using hydrodynamic waves

Demonstration of low-energy, wave-based information processing

Paving the way for fluid dynamic neuromorphic computing

Abstract

Nonlinear waves have played a historical role in laying the foundations of the science of complexity. Recently, they have also allowed the development of a new reservoir computing paradigm: neuromorphic computing by waves. In these systems, the information transmission acts as the excitation of wave dynamics, whose evolution processes the information to perform complex tasks at low energy consumption. To enable nonlinear hydrodynamic waves to do computing, we designed the Aqua-Photonic-Advantaged Computing Machine by Artificial Neural Networks (Aqua-PACMANN), a system where wave propagation in shallow water is the leading physical phenomenon, and the presence of electronics can be reduced to a CCD camera in detection. We show the realization of an XNOR logic gate as proof of concept of the Aqua-PACMANN's architecture and pave the way to a new class of fluid dynamic neuromorphic computing.