Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...

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

DY: Fachverband Dynamik und Statistische Physik

DY 18: Delay and Feedback Dynamics

DY 18.2: Vortrag

Montag, 16. März 2020, 17:15–17:30, ZEU 147

Analysing and Optimizing the Nonlinear Memory Capacity of Photonic Reservoir Computing — •Felix Köster and Kathy Lüdge — Institut für Theoretische Physik, TU Berlin, Hardenbergstraße 36, 10623 Berlin

Reservoir computing is a neuromorphic inspired machine learning paradigm that utilizes the naturally occurring computational capabilites of dynamical systems. In this work, we investigate the linear and nonlinear memory capacity of a delay-based class-A-laser reservoir computer via numerical simulations. We show that the reservoir computing performance is deeply connected to the total memory capacity and that resonances between the information injection rate and the delay time of the laser system play a crucial role in optimizing the reservoir. Additionally, we study the method of speed gradient descent as an optimization scheme for a delay based reservoir computer. By applying this method we can force our reservoir into having certain memory capacities tailored for a specific task.