Design, Implementation, Benchmarking of Hybrid Quantum-Classical Reservoir Computing

We report on (i) implementation choices, (ii) experimental results, (iii) comparison with simulations and alternative machine learning approaches, of echo state networks, with and without recurrence principles. We derive lessons learned, appropriate performance figures for ML (quantum or not) and expectations from tests on a variety of quantum hardware (analog and gate-model) - following the framework of hybriq quantum-classical reservoir computing (HQRC - arXiv:2311.14105) for time-series processing.