Summary
the 2014 International Symposium on Nonlinear Theory and its Applications
2014
Session Number:B2L-A
Session:
Number:B2L-A5
Towards on-chip implementation of fast delay-based reservoir computing using semiconductor ring lasers
Romain Modeste Nguimdo, Guy Verscha?elt, Jan Danckaert, Guy Van der Sande,
pp.308-311
Publication Date:2014/9/14
Online ISSN:2188-5079
DOI:10.34385/proc.46.B2L-A5
PDF download (116.7KB)
Summary:
In this contribution, we investigate the computational abilities of semiconductor ring lasers with optical delayed feedback for machine learning tasks. We benchmark our system on a chaotic time series prediction task. The results show that, as other types of lasers, good performance can be obtained in a broad range of the key parameters. In particular, we find that fast processing speed of 0:5 GSample/s can be achieved. This is obtained even for an overall delay loop of 2 ns. With this short delay, it is possible to implement this reservoir computing schemes on chip as both the semiconductor ring lasers and such a short delay can be easily implemented on the same chip.