A Chemical System that Recognizes the Shape of a Sphere

Abstract

Unconventional computing devices operating on nonlinear chemical media offer an interesting alternative to standard, semiconductor-based computers. In this work we consider database classifiers formed of interacting droplets in which a photosensitive variant of Belousov-Zhabotinsky (BZ) reaction proceeds. We introduce an evolutionary algorithm that searches for optimal construction of a droplets-based classifier for a given problem. The algorithm is based on maximizing the mutual information between the database and the observed evolution of medium. As an example application of chemical database classifiers we apply the idea to the dataset of points belonging to a unit cube. The dataset contains two output classes: 1 for points belonging to a sphere with radius 0.5 located in the cube center, and 0 for points outside of the sphere. The reliability of optimized chemical classifiers of such database for different numbers of droplets involved in data processing is presented.

Cancer classification with a network of chemical oscillators

“…The methods have been already used in a number of our recent papers and the technical details can be found there. 41,42 We study the time evolution of a network of droplets in the time interval [0,t sim ] and relate the output information to the number of excitations observed at a selected droplet within this time interval. The value of t sim is fixed and is not changed by the optimization procedure.…”

Applications of Information Theory Methods for Evolutionary Optimization of Chemical Computers

“…In this section, I briefly introduce oscillator networks, discuss the specific properties of the Belousov–Zhabotinsky reaction that can be useful for construction of classifiers and introduce a simple model of network time evolution. The detailed information can be found in [ 28 ].…”

“…However, the numerical complexity of models based on kinetic equations is still substantial, and they are too slow for large scale evolutionary optimization of a classifier made as a network of oscillators. Here, following [ 28 ], I use the event based model. I assume that three different phases: excited, refractive and responsive can be identified during a single oscillation cycle of a typical chemical oscillator [ 21 , 22 , 28 ].…”

“…Here, following [ 28 ], I use the event based model. I assume that three different phases: excited, refractive and responsive can be identified during a single oscillation cycle of a typical chemical oscillator [ 21 , 22 , 28 ]. The excited phase denotes the peak of activator concentration.…”

“…An oscillator in the response phase can get excited by interactions with an oscillator in the excited phase. Following the previous papers [ 28 ], the oscillation cycle combines the excitation phase lasting 1 s, the refractive phase, lasting 10 s and the responsive phase that is 19 s long (cf. Figure 2 ).…”