Search papers, labs, and topics across Lattice.
The paper introduces a kernel ridge regression framework for optimizing measurement operators in both stateless (QELMs) and stateful quantum reservoir computers (QRCs). This approach determines an optimal measurement operator that minimizes prediction error for a given reservoir and training data, offering improved efficiency compared to conventional QRC training, especially for large qubit numbers. The effectiveness of the method is demonstrated through numerical experiments on image classification and time series prediction.
Kernel ridge regression unlocks efficient training of quantum reservoir computers by optimizing measurement operators, outperforming conventional methods especially with large qubit counts.
Finding optimal measurement operators is crucial for the performance of quantum reservoir computers (QRCs), since they employ a fixed quantum feature map. We formulate the training of both stateless (quantum extreme learning machines, QELMs) and stateful (memory dependent) QRCs in the framework of kernel ridge regression. This approach renders an optimal measurement operator that minimizes prediction error for a given reservoir and training dataset. For large qubit numbers, this method is more efficient than the conventional training of QRCs. We discuss efficiency and practical implementation strategies, including Pauli basis decomposition and operator diagonalization, to adapt the optimal observable to hardware constraints. Numerical experiments on image classification and time series prediction tasks demonstrate the effectiveness of this approach, which can also be applied to other quantum ML models.
