Abstract
Via a Santa Fe time series prediction task, the performance of a reservoir computing (RC) system based on a delay feedback semiconductor laser (SL) under current modulation is experimentally investigated and evaluated in detail. Experimental results show that some key operation parameters including the modulation index, the feedback ratio and the SL bias current, seriously affect the prediction performance of this RC system. By optimizing parameter settings, the degradation of the RC performance with the increase of the SL bias current can be efficiently alleviated, and the RC system can still realize a good performance with a NMSE smaller than 0.1 even for the SL biased at a high level of three times threshold current Ith. This research is an improvement to a previously experimental report in which such RC system can achieve a good performance only under a SL bias level near I th .
Highlights
Nowadays, novel computational paradigms have attracted widespread attention due to the increasing demand for massive information processing [1], [2]
In 2011, Appeltant et al innovatively implemented a reservoir computing (RC) based on a simple nonlinear delay system [9], where the complex network is replaced by a single nonlinear element with delay feedback and the large number of nonlinear nodes are replaced by the virtual nodes distributed in a feedback loop
We focus on the variation of the system rest state with the feedback ratio k, where k is defined as a ratio of the feedback power and the output power of the DFB-semiconductor laser (SL) at free-running
Summary
Novel computational paradigms have attracted widespread attention due to the increasing demand for massive information processing [1], [2] Among these novel methods emerged in the past decades, recurrent neural network (RNN) is one of the most successfully neuro-inspired approaches. Properly training RNN is difficult due to the widely known problems of gradient vanishing and exploding [4], which partly limits its large scale deployment in practical application. This drawback can be overcome by reservoir computing (RC), which is originally known as echo state network and liquid.
Published Version (Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have