Year: 2024
Type: preprint
Abstract: This work explores the complexity and nonlinearity of seven different colloidal suspensions-Au, ferrofluid, TiO2}, ZnO, g-C3N4, MXene, and PEDOT:PSS-when electrically stimulated with fractal, chaotic, and random binary signals. The recorded electrical responses were analyzed using entropy, file compression, fractal dimension, and Fisher information measures to quantify complexity. The nonlinearity introduced by each colloid was evaluated by the deviation of the output from the best-fit hyperplane of the input-output mapping. The results showed that TiO2 was the most complex colloid across all inputs, exhibiting high entropy, poor compressibility, and an unpredictable response pattern. The colloids also exhibited significant nonlinearity, making them promising candidates for reservoir computation, where the mapping of inputs into high-dimensional nonlinear states is advantageous. This study provides insight into the dynamics of colloids and their potential for unconventional computational applications that exploit their inherent complexity and nonlinearity, and it provides a rapid method for assessing the suitability of a particular material for use as a computational substrate before others.
Source: arXiv (Cornell University)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4404313168