SKM 2023 – wissenschaftliches Programm

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

DY: Fachverband Dynamik und Statistische Physik

DY 46: Poster: Machine Learning and Data Analytics

DY 46.5: Poster

Donnerstag, 30. März 2023, 13:00–16:00, P1

Influence of mode-coupling on the information processing rate of Spin-VCSEL reservoir computer — •Lukas Mühlnickel, Lina Jaurigue, and Kathy Lüdge — Institut f. Physik, Technische Universität Ilmenau, Weimarer Str. 25, 98684 Ilmenau, Germany

The relative simplicity of reservoir computing, when comparing it to other machine learning methods, makes it suitable for efficient hardware implementation. The needed high dimensional reservoir dynamics can be provided by adding feedback to only one single nonlinear node, while driving the system with time multiplexed inputs. One promising realization utilizes the fast polarization dynamics of power efficient Spin-VCSELs. These fast field interactions are related to birefringence, dichroism and electron transition rates in the cavity material and occur on shorter time scales than the relaxation oscillations. Thus, compared to typical semiconductor lasers, much higher cutoff frequencies in the system response are observed for the Spin-VCSELs. We investigate the influence of these fast polarization oscillation dynamics on the reservoir performance when increasing data processing rates.