Summary
the 2014 International Symposium on Nonlinear Theory and its Applications
2014
Session Number:C2L-E
Session:
Number:C2L-E2
High-Speed Neuro-Inspired Information Processing Using Semiconductor Lasers
Ingo Fischer, Daniel Brunner, Miguel C. Soriano, Claudio R. Mirasso,
pp.549-552
Publication Date:2014/9/14
Online ISSN:2188-5079
Summary:
Photonic information processing has regained considerable interest due to the possibility to implement neuro-inspired concept in photonics hardware using delayed feedback systems. Employing induced transient dynamics of a semiconductor laser with delayed feedback, we can perform optical information processing at high data rates. Via time-multiplexing, we emulate a complex nonlinear photonic network using only a single or few lasers. Based on the induced dynamics of the network, classification tasks, time series prediction tasks, as well as vector and matrix operations can be implemented. Excellent performance can be achieved injecting the information all-optically at rates up to several GSamples/s.