Expressive limits of quantum reservoir computing

Quantum machine learning (QML) is a strongly growing research field, dominated by ansatzes that use parameterized quantum gates. An alternative QML paradigm is quantum reservoir computing (QRC) that originates from artificial neural networks. We discuss the close con- nection between both approaches and demonstrate a realization of the transverse-field Ising model on a gate-based quantum computer. Advantages of QRC do arise from the large state-space dimension and internal degrees of freedom that can exceed capabilities of current noisy intermediate-scale quantum (NISQ) platforms.