Key points are not available for this paper at this time.
We numerically study how to enhance reservoir computing performance by thoroughly extracting the spin-wave device potential for higher-dimensional information generation. The reservoir device has a 1-input exciter and 120-output detectors on the top of a continuous magnetic garnet film for spin-wave transmission. For various nonlinear and fading-memory dynamic phenomena distributing in the film space, small in-plane magnetic fields are used to prepare stripe domain structures and various damping constants at the film sides and bottom are explored. The ferromagnetic resonant frequency and relaxation time of spin precession clearly characterizes the change in spin dynamics with the magnetic field and damping constant. The common input signal for reservoir computing is a 1-GHz cosine wave with random 6-valued amplitude modulation. A basic 120-dimensional reservoir output vector is obtained from time-series signals at the 120-output detectors under each of three magnetic field conditions. Then, 240- and 360-dimensional reservoir output vectors are also constructed by concatenating two and three basic ones, respectively. In nonlinear autoregressive moving average (NARMA) prediction tasks, the computational performance is enhanced as the dimension of the reservoir output vector becomes higher and a significantly low prediction error is achieved for the tenth-order NARMA task using the 360-dimensional vector and optimum damping constant. The results are clear evidence that the collection of diverse output signals efficiently increases the dimensionality of the integrated reservoir state vector (i.e. reservoir-state richness) and thereby contributes to high computational performance. This paper demonstrates that performance enhancement through various configuration settings is a practical approach for on-chip reservoir computing devices with small numbers of real output nodes.
Nakane et al. (Tue,) studied this question.