Photonic information processing at 20GS/s rates based on semiconductor lasers with delayed optical feedback
Bueno, Julián; Brunner, Daniel; Soriano, Miguel C.; Fischer, Ingo
2017 Lasers and Electro-Optics Europe & European Quantum Electronics Conference (CLEO/Europe-EQEC), IEEE, (2017)
Photonic implementations of novel information processing schemes based on machine learning approaches like reservoir computing (RC)  have raised strong interest in the last years due to their speed and performance. A reservoir computer, conceptionally as simple as a semiconductor laser with delayed optical feedback already proved excellent performance at 5 GSamples/s rates thanks to its nonlinear responses in the GHz regime . Reported bandwidths of 80GHz of semiconductor laser with optical injection  open the possibility of realizing photonic reservoir computers based on semiconductor lasers at high speeds. The study of the properties of the system and their impact on the RC performance is crucial to better understand the underlying physics, improve designs and tailor systems for specific tasks.