RESERVOIR COMPUTING: USING CONJUGATED POLYMERS FOR CONSTRUCTING PHYSICAL RESERVOIRS

Rafael Francisco Santiago de Souza; Bruno Bassi Millan Torres; Gregório Faria

All Papers

Submission

Paper Title

RESERVOIR COMPUTING: USING CONJUGATED POLYMERS FOR CONSTRUCTING PHYSICAL RESERVOIRS

Authors

Rafael Francisco Santiago de Souza
Bruno Bassi Millan Torres
Gregório Faria

Modality

Pôster

Subject area

Síntese e caracterização de materiais

Publishing Date

12/12/2023

Country of Publishing

Brasil

Language of Publishing

Inglês

Paper Page

https://www.even3.com.br/anais/workshop-do-ineo-2023/613066-reservoir-computing--using-conjugated-polymers-for-constructing-physical-reservoirs

ISBN

978-65-272-0088-8

Keywords

Reservoir, computing, polymer,physical

Summary

Reservoir computing is a novel idea that emerged in the early 2000s. It involves utilizing the nonlinear dynamics of physical systems, known as reservoirs, as a platform to build neural networks. Unlike classical neural networks, where all elements require pre-training, the non-classical reservoir operates without prior training. The reservoir leverages the nonlinear dynamics of its elements to transmit information, and only the outlet layer of the reservoir is trained for a desired regression or classification [1]. The work presented here explores the creation of a physical reservoir using electropolymerized conducting polymer microfibers based on EDOT, pyrrole, and aniline derivatives. It was found that PEDOT microfibersbased on EDOT, pyrrole, and aniline derivatives. It was found that EDOT microfibers were easy to grow using a sinusoidal voltage, with amplitudes between 2 and 10 V and frequencies ranging from 25 to 200 Hz. In contrast, no growth has occurred in any of the tested conditions for Aniline, o-Anisidine and pyrrole. The authors then built a physical reservoir by placing a monomeric solution in contact with several gold electrodes and applying a sinusoidal voltage to enable polymeric fibers to grow and connect to the electrodes. The physical reservoirs obtained using PEDOT were evaluated in regards to their dynamical responses, fading memory, and nonlinear transformations[3]. In addition, a single fiber of PEDOT was grown in patterned golden substrates, allowing the authors to use the fiber as an Organic Electrochemical Transistor (OECT) [2], extracting information about their output curves. These classical characterization curves used to measure performance in OECTs were used to withdraw information about the non-linear behavior coming from a single fiber, which is the physical unit within the reservoir. So far, preliminary tests allowed one to understand the reservoir’s dynamics, gather physical parameters, and adjust them in order to optimize the proposed reservoir. [1] Nakajima, Kohei & Hauser, Helmut & Li, Tao & Pfeifer, Rolf. (2015). Information processing via physical soft body. Scientific reports. 5. 10487. 10.1038/srep10487. [2] Rivnay, J., Inal, S., Salleo, A. et al. Organic electrochemical transistors. Nat Rev Mater 3, 17086 (2018). https://doi.org/10.1038/natrevmats.2017.86 [3] Tanaka, Gouhei & Yamane, Toshiyuki & Heroux, Jean & Nakane, Ryosho & Kanazawa, Naoki & Takeda, Seiji & Numata, Hidetoshi & Nakano, Daiju & Hirose, Akira. (2019). Recent advances in physical reservoir computing: A review. Neural Networks. 115. 10.1016/j.neunet.2019.03.005.

Title of the Event: Workshop do INEO 2023
City of the Event: Nazaré Paulista
Title of the Proceedings of the event: Anais do Workshop INEO 2023
Name of the Publisher: Even3
Means of Dissemination: Meio Digital

How to cite

SOUZA, Rafael Francisco Santiago de; TORRES, Bruno Bassi Millan; FARIA, Gregório. RESERVOIR COMPUTING: USING CONJUGATED POLYMERS FOR CONSTRUCTING PHYSICAL RESERVOIRS.. In: Anais do workshop INEO 2023. Anais...Nazaré Paulista(SP) Hotel Estância Atibainha, 2023. Available in: https//www.even3.com.br/anais/workshop-do-ineo-2023/613066-RESERVOIR-COMPUTING--USING-CONJUGATED-POLYMERS-FOR-CONSTRUCTING-PHYSICAL-RESERVOIRS. Access in: 25/04/2025