Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 39: Machine Learning in Dynamics and Statistical Physics II

DY 39.3: Vortrag

Donnerstag, 20. März 2025, 15:30–15:45, H47

Physical Reservoir Computing with Ferroelectric Oxides — •Atreya Majumdar¹, Yan Meng Chong², Dennis Meier², and Karin Everschor-Sitte¹ — ¹Faculty of Physics and Center for Nanointegration Duisburg-Essen (CENIDE), University of Duisburg-Essen, Duisburg, Germany — ²Department of Materials Science and Engineering, Norwegian University of Science and Technology (NTNU), Trondheim, Norway

Physical reservoir computing has shown remarkable potential in magnetic systems by utilizing their complex, non-linear, and history-dependent intrinsic dynamics for machine learning tasks [1]. More recently, ferroelectric materials - the electrical analogs of magnetic systems - have garnered attention. These materials not only meet all the essential criteria for reservoir computing but also bring unique advantages [2]. Here, we introduce the ferroelectric semiconductor ErMnO3 as a novel physical reservoir. By utilizing the material's non-linear and history-dependent photocurrent response, we demonstrate its capability to recognize varying input light pulse intensities. This study highlights the potential of ferroelectric materials in physical reservoir computing, paving the way for energy-efficient and scalable computing architectures.

[1] O. Lee, et al., Perspective on unconventional computing using magnetic skyrmions. Appl. Phys. Lett. 122, 260501 (2023).

[2] K. Everschor-Sitte, A. Majumdar, et al., Topological magnetic and ferroelectric systems for reservoir computing. Nat. Rev. Phys. 6, 455 (2024).

Keywords: Physical reservoir computing; Machine learning; Pattern recognition; Ferroelectric materials; Magnetic materials