direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments

Neuromorphic Computing using QD-Networks

Schematic representation of reservoir computing based on an optical network.

The main objective of this project is to implement Reservoir Computing (RC, a neuro-inspired information processing scheme) in an optical network of nano-structures. Its realization requires spectrally tailored quantum dot micropillar arrays (QDMPA) and diffractive coupling to establish all-optical networks including hundreds of such emitters. Our underlying interdisciplinary approach combines three recent concepts by bridging nanostructures to a macroscopic complex system which is utilized for powerful computation. Namely, these concepts are RC as the functional concept, QDMPAs as the hardware platform, and diffractive coupling schemes for scalable optical networks to implement the complex neuro-inspired systems, capable of ultra-high speed information processing. It represents a unique opportunity to integrate these three concepts into a fully functional computing system with great potential in terms of performance, speed, compactness, energy-efficiency and future extensions to quantum machine learning.


More information: http://neuroqnet.com/



Dr. Daniel Brunner, Department of Optics, FEMTO-ST, Besançon, France

Funded by:

Volkswagen Stiftung


A. Kaganskiy et al., Enhancing the photon-extraction efficiency of site-controlled quantum dots by deterministically fabricated microlenses, ArXiv e-prints, 1708.03512 (2017)

M. Strauß et al., Resonance fluorescence of a site-controlled quantum dot realized by the buriedstressor growth technique, Appl. Phys. Lett. 110, 111101 (2017)


Zusatzinformationen / Extras


Schnellnavigation zur Seite über Nummerneingabe