Neutral-atom quantum devices offer a natural and powerful platform for Quantum Reservoir Computing (QRC), a non-variational approach to quantum-enhanced machine learning that is attracting growing interest in the Noisy Intermediate-Scale Quantum (NISQ) era. In this work, we implement a QRC-based learning model tailored to Pasqal neutral-atom hardware, where classical data are encoded into time-dependent detuning waveforms and processed through programmable many-body dynamics. Feature embeddings are extracted from single- and two-qubit Pauli-$Z$ observables and used within a simple linear readout model. We systematically investigate both global and local detuning encoding strategies on a supervised image classification task, benchmarking performance against a purely classical linear baseline. Our results show that processing the data through the quantum reservoir yields a consistent $\sim$4\% improvement in accuracy, highlighting the system effectiveness as a nonlinear feature map. Both encoding strategies exhibit comparable performance, suggesting that QRC can leverage different control schemes while remaining compatible with near-term and emerging hardware capabilities. These findings provide concrete evidence that QRC can enhance classical machine learning pipelines by injecting non-classical correlations without the need for variational optimization, thereby avoiding issues such as barren plateaus and training instabilities. More broadly, this work establishes neutral-atom platforms as a promising and scalable setting for QRC, opening the way toward practical quantum-enhanced learning on near-term devices.
Neutral-atom quantum devices offer a natural and powerful platform for Quantum Reservoir Computing (QRC), a non-variational approach to quantum-enhanced machine learning that is attracting growing interest in the Noisy Intermediate-Scale Quantum (NISQ) era. In this work, we implement a QRC-based learning model tailored to Pasqal neutral-atom hardware, where classical data are encoded into time-dependent detuning waveforms and processed through programmable many-body dynamics. Feature embeddings are extracted from single- and two-qubit Pauli-Z observables and used within a simple linear readout model. We systematically investigate both global and local detuning encoding strategies on a supervised image classification task, benchmarking performance against a purely classical linear baseline. Our results show that processing the data through the quantum reservoir yields a consistent ∼4% improvement in accuracy, highlighting the system effectiveness as a nonlinear feature map. Both encoding strategies exhibit comparable performance, suggesting that QRC can leverage different control schemes while remaining compatible with near-term and emerging hardware capabilities. These findings provide concrete evidence that QRC can enhance classical machine learning pipelines by injecting non-classical correlations without the need for variational optimization, thereby avoiding issues such as barren plateaus and training instabilities. More broadly, this work establishes neutral-atom platforms as a promising and scalable setting for QRC, opening the way toward practical quantum-enhanced learning on near-term devices.
In this thesis we investigate Quantum Reservoir Computing (QRC) as a non-variational paradigm for quantum machine learning on near-term neutral-atom devices. Motivated by the limitations of NISQ hardware, we implement a QRC pipeline in Pasqal’s Pulser environment that encodes classical inputs into detuning waveforms and extracts embeddings from Pauli-Z and ZZ observables. We design two encoding strategies: a global detuning applied to all atoms and a local detuning. Their performance is benchmarked against a classical PCA baseline using logistic regression. We find that QRC embeddings consistently outperform the classical features, with peak test accuracies of 90.17% (global) and 89.61% (local) compared to ≃ 85% for PCA. The advantage is especially marked in low-data regimes, where quantum dynamics enrich the feature space accessible to linear classifiers. We also identify practical challenges, such as symmetry-induced redundancy in global encoding and the absence of per-qubit detuning in current hardware. This establishes QRC as a promising and experimentally grounded framework to exploit the dynamical richness of neutral-atom platforms for supervised learning in the NISQ era.
Quantum Reservoir Computing (QRC) leverages the natural dynamics of quantum systems for information processing, without requiring a fault-tolerant quantum computer. In this work, we apply QRC within a hybrid quantum classical framework for image denoising. The quantum reservoir is implemented using a Rydberg atom array, while a classical neural network serves as the readout layer. To prepare the input, images are first compressed using Principal Component Analysis (PCA), reducing their dimensionality to match the size of the atom array. Each feature vector is encoded into local detuning parameters of a time-dependent Hamiltonian governing the Rydberg system. As the system evolves, it generates nonlinear embeddings through the measurement of observables across multiple time steps. These temporal embeddings capture complex correlations, which are fed into a classical neural network to reconstruct the denoised images. To evaluate performance, we compare this QRC-assisted model against a baseline architecture consisting of PCA followed by a dense neural network, trained under identical conditions. Our results show that the QRC-based approach achieves improved image sharpness and similar structural recovery compared to the PCA-based model. We demonstrate the practical viability of this framework through experiments on QuEra's Aquila neutral-atom processor, leveraging its programmable atom arrays to physically realize the reservoir dynamics.
Quantum reservoir computing (QRC) uses the dynamics of a fixed or weakly tuned quantum system to transform temporal and sequential inputs into measured features, while training is typically confined to a classical readout. This separation reduces reliance on repeated quantum parameter updates and avoids the barren plateaus associated with variational circuit training. Its computational power is often attributed to the exponentially large Hilbert space of the quantum system. However, the memory, nonlinearity, and expressivity that determine what a reservoir can actually compute depend jointly on the input encoding, quantum evolution, observables, measurement, and readout, not on Hilbert space dimension alone. On hardware, these capabilities are further constrained by finite sampling, hardware noise, measurement backaction, and the cost of estimating observables, so a large state space alone does not guarantee useful computation. In this survey, we develop a common system model that connects these components and use it to organize QRC foundations, computational properties, reservoir architectures, operating protocols, and physical implementations. We examine spin, photonic, superconducting, bosonic, neutral atom, and other analog platforms, together with applications, software and high performance computing support, benchmarking, and reproducibility. The analysis distinguishes hardware demonstrations from simulations and identifies the assumptions and resources that govern comparisons across implementations. Current results do not establish a broad quantum advantage over well matched classical reservoirs. We therefore specify the resource accounting, benchmark standards, and theoretical criteria needed to evaluate claims of quantum advantage.
In this paper, we define and benchmark a hybrid quantum-classical machine learning pipeline by performing a binary classification task applied to a real-world financial use case. Specifically, we implement a Quantum Reservoir Computing (QRC) layer within a classical routine that includes data preprocessing and binary classification. The reservoir layer has been executed on QuEra's Aquila, a 256-qubit neutral atom simulator, using two different types of encoding: position and local detuning. In the former case, classical data are encoded into the relative distance between atoms; in the latter, into pulse amplitudes. The developed pipeline is applied to predict credit card defaults using a public dataset and a wide variety of traditional classifiers. The results are compared with a fully-classical pipeline including a Deep Neural Network (DNN) model. Additionally, the impact of hardware noise on classification performance is evaluated by comparing the results obtained using Aquila within the classification workflow with those obtained using a classical, noiseless emulation of the quantum system. The results indicate that the noiseless emulation achieves competitive performance with the fully-classical pipeline, while noise significantly degrades overall performance. Although the results for this specific use case are comparable to those of the classical benchmark, the flexibility and scalability of QRC highlight strong potential for a wide range of applications.
Quantum Reservoir Computing (QRC) has emerged as a powerhouse for Noisy Intermediate-Scale Quantum (NISQ) machine learning, offering the ability to process complex temporal data with minimal training overhead by leveraging the high-dimensional dynamics of quantum states. This paper introduces QRC-Lab, an open-source, modular Python framework designed specifically to facilitate the transition from theoretical quantum dynamics to applied machine learning research. We provide a rigorous definition of QRC, contrasting physical and gate-based approaches, and detail the mathematical foundations of the reservoir mapping. QRC-Lab’s gate-based architecture is presented as a flexible laboratory for exploring data encoding, reservoir connectivity, and measurement strategies. We validate the framework through diverse case studies, including Short-Term Memory capacity, non-linear Parity check, and NARMA-type forecasting. Furthermore, we discuss the framework’s future roadmap towards multi-platform hardware support and its role in visualizing theoretical risk bounds, positioning QRC-Lab as a vital tool for the next generation of quantum computing education.
Biomarker-based prediction of clinical outcomes is challenging due to nonlinear relationships, correlated features, and the limited size of many medical datasets. Classical machine-learning methods can struggle under these conditions, motivating the search for alternatives. In this work, we investigate quantum reservoir computing (QRC), using both noiseless emulation and hardware execution on the neutral-atom Rydberg processor \textit{Aquila}. We evaluate performance with six classical machine-learning models and use SHAP to generate feature subsets. We find that models trained on emulated quantum features achieve mean test accuracies comparable to those trained on classical features, but have higher training accuracies and greater variability over data splits, consistent with overfitting. When comparing hardware execution of QRC to noiseless emulation, the models are more robust over different data splits and often exhibit statistically significant improvements in mean test accuracy. This combination of improved accuracy and increased stability is suggestive of a regularising effect induced by hardware execution. To investigate the origin of this behaviour, we examine the statistical differences between hardware and emulated quantum feature distributions. We find that hardware execution applies a structured, time-dependent transformation characterised by compression toward the mean and a progressive reduction in mutual information relative to emulation.