Impact of input mask signals on delay-based photonic reservoir computing with semiconductor lasers

Abstract

We experimentally investigate delay-based photonic reservoir computing using semiconductor lasers with optical feedback and injection. We apply different types of temporal mask signals, such as digital, chaos, and colored-noise mask signals, as the weights between the input signal and the virtual nodes in the reservoir. We evaluate the performance of reservoir computing by using a time-series prediction task for the different mask signals. The chaos mask signal shows superior performance than that of the digital mask signals. However, similar prediction errors can be achieved for the chaos and colored-noise mask signals. Mask signals with larger amplitudes result in better performance for all mask signals in the range of the amplitude accessible in our experiment. The performance of reservoir computing is strongly dependent on the cut-off frequency of the colored-noise mask signals, which is related to the resonance of the relaxation oscillation frequency of the laser used as the reservoir.

Recent advances in physical reservoir computing: A review

“…Performance improvement was observed, provided that the mask frequency was near the relaxation oscillation of the ring cavity laser. It was experimentally confirmed that a chaotic or colored-noise mask gives better results with a properly selected cut-off frequency (Kuriki et al, 2018). In another study, numerical simulation of a reservoir in which the input signal was applied to a tunable Bragg reflector providing external feedback to the laser cavity was proposed, and the reservoir was applied to a waveform classification task .…”

Performance boost of time-delay reservoir computing by non-resonant clock cycle

“…which is piecewise constant with step length θ = τ /N and values w n . Multiple options for the choice of a mask function are compared in reference [19]. The final preprocessed input signal J(t) is…”

Microfabricated Neuroaccelerometer: Integrating Sensing and Reservoir Computing in MEMS

“…This led to an interconnection scheme where each virtual node was coupled at least to one of its nearest neighbor: the state of a given node depended on the state of the previous node, a time θ in the past. While this mask signal could take different forms, such as different random distributions [3], [19], sinusoidal signals [20], multiple level digital signals [21] and even chaotic signals [22], we chose the most commonly used form of a random binary sequence for the sake of simplicity. In this case, the mask was composed of only two values and switched randomly between them after each interval θ.…”