Reservoir Computing with Colloidal Mixtures of ZnO and Proteinoids

Abstract

Liquid computers use incompressible fluids for computational processes. Here we present experimental laboratory prototypes of liquid computers using colloids composed of zinc oxide (ZnO) nanoparticles and microspheres containing thermal proteins (proteinoids). The choice of proteinoids is based on their distinctive neuron-like electrical behaviour and their similarity to protocells. In addition, ZnO nanoparticles are chosen for their non-trivial electrical properties. Our research demonstrates the successful extraction of 2-, 4- and 8-bit logic functions in ZnO proteinoid colloids. Our analysis shows that each material has a distinct set of logic functions, and that the complexity of the expressions is directly related to each material present in a mixture. These findings provide a basis for the development of future hybris liquid devices capable of general purpose computing.