EUVIMEDEUVIMED entdecken
Crossref

Dynamic Control of Weight-Update Linearity in Magneto-Ionic Synapses

Guillaume Bernard, Kellian Cottart, Maria-Andromachi Syskaki, Victor Porée, Andrea Resta iD, Alessandro Nicolaou, Alan Durnez, Shimpei Ono, Ariam Mora Hernandez, Juergen Langer, Damien Querlioz, Liza Herrera Diez iD

Nano Letters · 29.01.2025 · 25 (4) · 1443-1450

DOI: 10.1021/acs.nanolett.4c05247 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Nano Letters, online erschienen am 13.01.2025. Verfasst von Guillaume Bernard und 11 weiteren Autor:innen. Beteiligte Einrichtung: Centre de Nanosciences et de Nanotechnologies, CNRS, Université Paris-Saclay, 91120 Palaiseau, France. Hauptthemen laut MEDLINE: Synapses, Neural Networks, Computer. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 56 Literaturangaben. Gefördert von: Agence Nationale de la Recherche, HORIZON EUROPE European Innovation Council, French RENATECH network.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multifunctional hardware technologies for neuromorphic computing are essential for replicating the complexity of biological neural systems, thereby improving the performance of artificial synapses and neurons. Integrating ionic and spintronic technologies offers new degrees of freedom to modulate synaptic potentiation and depression, introducing novel magnetic functionalities alongside the established ionic analogue behavior. We demonstrate that magneto-ionic devices can perform as synaptic elements with dynamically tunable depression linearity controlled by an external magnetic field, a functionality reminiscent of neuromodulation in biological systems. By applying magnetic fields we significantly reduce the nonlinearity of synaptic depression, transitioning from an exponential dependence to a linear response at higher fields. Neural network simulations reveal that this magnetically induced linearity enhancement improves learning accuracy across a wide range of learning rates, which is retained after the magnetic field is removed. These findings highlight the versatility and promise of magneto-ionic devices for developing tunable synaptic elements for neuromorphic hardware.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

SynapsesNeural Networks, ComputerMagnetic FieldsNeurons

Stichwörter der Autor:innen

artificial synapsesmagneto-ionicsnanodevicesweight-update linearity

Autor:innen und Institutionen

  • Centre de Nanosciences et de Nanotechnologies, CNRS, Université Paris-Saclay, 91120 Palaiseau, France.
  • Singulus Technology AG, Hanauer Landstrasse 103, 63796 Kahl am Main, Germany.
  • Synchrotron SOLEIL, L'Orme des Merisiers, 91190 Saint-Aubin, France.
  • International Center for Synchrotron Radiation Innovation Smart, Tohoku University, Aoba-Ku, Sendai 980-8572, Japan.

ORCID-Profile: Andrea Resta, Liza Herrera Diez

Förderung

  • Agence Nationale de la Recherche
  • HORIZON EUROPE European Innovation Council
  • French RENATECH network

Zitieren

APA

Bernard, G., Cottart, K., Syskaki, M. A., Porée, V., Resta, A., Nicolaou, A., Durnez, A., Ono, S., Mora Hernandez, A., Langer, J., Querlioz, D., & Herrera Diez, L. (2025). Dynamic Control of Weight-Update Linearity in Magneto-Ionic Synapses. Nano Letters, 25(4), 1443-1450. https://doi.org/10.1021/acs.nanolett.4c05247

Vancouver

Bernard G, Cottart K, Syskaki MA, Porée V, Resta A, Nicolaou A, et al. Dynamic Control of Weight-Update Linearity in Magneto-Ionic Synapses. Nano Letters. 2025;25(4):1443-1450. doi:10.1021/acs.nanolett.4c05247

BibTeX
@article{Bernard2025,
  author = {Bernard, Guillaume and Cottart, Kellian and Syskaki, Maria-Andromachi and Porée, Victor and Resta, Andrea and Nicolaou, Alessandro and Durnez, Alan and Ono, Shimpei and Mora Hernandez, Ariam and Langer, Juergen and Querlioz, Damien and Herrera Diez, Liza},
  title = {Dynamic Control of Weight-Update Linearity in Magneto-Ionic Synapses},
  journal = {Nano Letters},
  year = {2025},
  volume = {25},
  number = {4},
  pages = {1443--1450},
  doi = {10.1021/acs.nanolett.4c05247}
}
Kostenlos starten →