Handwritten digit recognition using a biologically-inspired physical reservoir of single-transistor chaotic oscillators

Year: 2026

Authors: Fossi JT., Xie DH., Zhao MY., Tegnitsap JVN., Concas R., Frasca M., Meucci R., Minati L.

Autors Affiliation: Univ Elect Sci & Technol China, Sch Life Sci & Technol, Chengdu 611731, Peoples R China; Univ Dschang, Fotso Victor Univ Inst Technol, POB 134, Dschang, Cameroon; Univ Dschang, Dept Phys, Res Unit Condensed Matter Elect & Signal Proc, POB 67, Dschang, Cameroon; Ist Nazl Ric Metrolog INRiM, I-50019 Sesto Fiorentino, Italy; European Lab Nonlinear Spect LENS, I-50019 Sesto Fiorentino, Italy; Univ Catania, Dept Elect Elect & Comp Engn, I-95131 Catania, Italy; Univ Florence UNIFI, Dept Phys, I-50019 Sesto Fiorentino, Italy; Ist Nazl Ott CNR INO, I-50125 Florence, Italy; Univ Trento, Dept Phys, I-38123 Trento, Italy; Inst Sci Tokyo, Inst Innovat Res, Nano Sensing Res Unit, Yokohama 2268503, Japan.

Abstract: In this paper, we investigate handwritten-digit recognition using a compact physical reservoir computer based on a network of 59 single-transistor chaotic oscillators derived from a neuronal culture. Input images from the MNIST dataset are first reduced to 8 & times; 8 pixels and then encoded as node-dependent control voltages applied locally to the nodes of the oscillator network. A mapping strategy in which the most informative pixels, identified based on the variance or ANOVA F-value, are assigned preferentially to structurally central nodes of the reservoir network, is employed. These static images are transformed within the network into class-dependent spatiotemporal responses from which low-dimensional parameters, including temporal standard deviation, mean phase synchronization, and spectral measures, are extracted. The obtained features are subsequently processed by a shallow multilayer perceptron acting as the readout stage. Numerical simulations and physical experiments with an analog electronic setup show that the reservoir achieves its best performance in an intermediate coupling regime, where sufficient coordination coexists with dynamical heterogeneity. Under these conditions, the classification accuracy obtained from the numerical simulations reaches approximately 81%; significantly higher than the baseline obtained using only image pixels with the selected smallscale classification network. From experimental measurements on the circuit board, the best accuracy reaches 75% and is obtained when temporal variability parameters are combined with the raw image representation. Beyond classification performance, this study demonstrates that a biologically inspired topology can play a central role in small-scale physical reservoirs by determining how information is distributed, transformed, and separated through collective nonlinear dynamics. This study thus establishes chaotic electronic oscillator networks as viable and interpretable analog substrates for reservoir computing.

Journal/Review: CHAOS SOLITONS & FRACTALS

Volume: 212      Pages from: 118874-1  to: 118874-22

More Information: L.M. gratefully acknowledges the support of the Hundred Talents program of the University of Electronic Science and Technology of China, of the Outstanding Young Talents Program (Overseas) program of the National Natural Science Foundation of China, and of the talent programs of the Sichuan province and Chengdu municipality. M.Z. contributed to the study over the period Jan-Oct 2025. All experimental activities were fully self-funded and conducted by L.M. using own independent assets located in Grigno TN, Italy; the manuscript was later finalized after joining the University of Electronic Science and Technology of China, to which no equipment was transferred.
KeyWords: Physical reservoir computing; Chaotic electronic oscillators; Structural connectivity; Handwritten digit recognition; Multilayer perceptron
DOI: 10.1016/j.chaos.2026.118874