Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Frontiers in Quantum Science and Technology

Convolutional quantum extreme learning machines for time series forecasting: a case study on solar energy generation

Francesco Strata, Luca Migliori, Emanuele Broggini, Sara Pezzuolo, Cecilia Maria Ortoleva, Emiliano Luzietti

Published Nov 9, 2026• 0 citations
PDF (CC BY)Publisher

Abstract

Introduction Accurate forecasting of solar power generation is essential for grid stability, yet it remains difficult because of the strong volatility and non-linear dynamics of solar resources. Classical machine learning models have improved forecasting performance, but they often fail to capture rapid changes in generation. Methods We propose a hybrid forecasting framework, Convolutional Quantum Extreme Learning Machines (C-QELM), which combines a fixed gate-based quantum feature map with local time-series patches and uses the resulting high-dimensional representations in classical forecasters. To evaluate the approach, we conducted a controlled synthetic benchmark comparing C-QELM with a global Standard QELM, a convolutional purely classical control, and Random Fourier Feature (RFF) maps. We also performed rolling one-step-ahead experiments on three solar datasets, comparing C-QELM with an unaugmented baseline and a dimension-matched classical convolutional random-feature control (CRC). Results In the synthetic benchmark, C-QELM substantially outperformed the global QELM and achieved the lowest mean error among the non-task-informed random-feature methods, although its confidence intervals overlapped with those of CRC and the best RFF configuration. Results on the solar datasets were similarly mixed, indicating that performance depends on both the dataset and the downstream architecture. Discussion Overall, C-QELM does not provide an unconditional performance advantage, but it may reveal a conditional benefit in highly dynamic regimes. Its small register size, low circuit depth, patch-wise reset, and lack of quantum-parameter training make it a promising architecture for future testing on NISQ platforms.

DOI
10.3389/frqst.2026.1892471

Related papers

Quantum Machine Learning for Solar Power Generation Forecasting

Abdul Rehman, Muhammad Aurangzeb, Sheeraz Iqbal +2 more

Accurate forecasting of solar power generation plays a vital role in maximising energy output and ensuring grid stability in renewable energy systems. Conventional methods of forecasting can have a hard time modelling the intricate interactions of dynamic environmental factors, like variations in temperature, solar irradiance, and time-of-day effects. This paper proposes a Quantum Machine Learning (QML)-based model that is specifically focused on solar power forecasting and combines three approaches: Quantum Kernel Learning (QKL), K-Nearest Neighbours (KNN) and Linear Regression (LR). Using quantum-enhanced feature mapping and optimised quantum kernels, the model is able to deal with highly non-linear dependencies that are not easily modelled by classical approaches. The proposed approach is named the Hybrid Quantum Classical Model (HQCM) and compared to the classical KNN and LR models in terms of performance parameters. Training and testing were done using data obtained from solar power generation. The simulated results proved that the proposed HQCM can reduce Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) by 57.2% and 55.8%, respectively, with high stability and reliability predictions. The proposed model has achieved a good$R^{2}$of 0.999 with the Quantum K-Nearest Neighbours (QKNN) model and has an absolute error as high as 1.4 MW when the test dataset is recorded. Cumulative Distribution Function (CDF) and residual distribution plots affirm that the quantum model errors are concentrated close to zero, while classical baselines exhibit much larger residual magnitudes. The proposed HQCM has the potential to be helpful in managing the energy grid and making the integration of solar power into smart grids easier. Future work will build on this to real-time optimisation of power grids while considering how other renewable energy sources can be optimised in this way.

International Conference on Electrical, Computer and Energy Technologies (ICECET)•Jul 1, 2026•0 citations

Optimizing Hybrid Quantum-Classical Models Using Parameterized Quantum Circuits for Time-Series Forecasting

Harshraj Bhoite

Time-series forecasting is vital across sectors such as finance, energy, and environmental science, where accurate future predictions drive informed decision-making. Traditional statistical methods and deep learning models like LSTM and Transformers, while effective, often struggle with high-dimensional data and long-term dependencies. In this paper, we propose a hybrid quantum-classical model that utilizes Parameterized Quantum Circuits (PQCs) to capture complex temporal patterns in multivariate time-series data. Our architecture integrates a classical encoder, a quantum variational processor, and a classical decoder, trained end-to-end using gradient-based optimization. Experiments conducted on benchmark datasets—electricity load, stock prices, and weather forecasting—demonstrate that the hybrid model consistently outperforms classical baselines in terms of Mean Squared Error (MSE), Mean Absolute Error (MAE), and R² score. Robustness analysis confirms the model’s resilience to moderate quantum noise, while an ablation study highlights the performance gain from quantum components. This work showcases the potential of hybrid quantum-classical approaches on Noisy Intermediate-Scale Quantum (NISQ) devices and lays a foundation for scalable, real-world quantum machine learning applications.

•Jul 19, 2025•0 citations

Entanglement and classical simulability in quantum extreme-learning machines

A. De Lorenzis, M. P. Casado, Nicola Lo Gullo +3 more

Quantum Machine Learning (QML) has emerged as a promising framework for exploring how quantum dynamics may enhance data processing tasks. Here we investigate Quantum Extreme Learning Machines (QELMs), a quantum analogue of classical Extreme Learning Machines in which training is restricted to the output layer. Our architecture combines dimensionality reduction (via PCA or Autoencoders), quantum state encoding, evolution under an XX Hamiltonian, and projective measurement to produce features for a classical single-layer classifier. By analyzing the classification accuracy as a function of evolution time, we observe a sharp transition between low- and high-accuracy regimes, followed by saturation. Remarkably, the saturated performance is comparable to that obtained using Haar-random unitaries that generate maximally complex dynamics, even though the XX model is integrable and local. Our results indicate that this increase in performance correlates with the onset of entanglement, which improves the embedding of classical data in Hilbert space and leads to more separable clusters in measurement probability space. Thus, moderate entanglement can contribute positively to the structure of the data representation, improving learnability without necessarily implying quantum computational advantage. For the image-classification tasks studied here, namely MNIST, Fashion-MNIST, and CIFAR-10, the relevant evolution time is consistent with information exchange over short distances and, within the explored system sizes, does not show evidence of scaling with the full system size. This suggests that QELM performance in this regime relies only on limited entanglement and remains compatible with efficient classical simulation. Our results clarify how local quantum dynamics and moderate quantum correlations are already sufficient to generate useful feature representations for learning.

arXiv:2509.06873•Physical Review Applied•Apr 23, 2026•9 citations

Breaking concentration barriers for quantum extreme learning on digital quantum processors

Timothée Dao, Ege Yilmaz, Ibrahim Shehzad +8 more

Reservoir computing leverages rich, non-linear dynamics to process temporal data. Quantum variants promise enhanced expressivity from high-dimensional Hilbert spaces, yet their practical applicability is hindered by hardware noise and concentration effects that can erase input-output distinguishability at large system sizes. In this work, we present and experimentally demonstrate a Quantum Extreme Learning Machine (QELM) tailored to state-of-the-art superconducting platforms, employing up to 124 qubits and circuits with more than 5,000 two-qubit gates on IBM Quantum computers. We introduce a practical multi-objective hyperparameter tuning strategy that jointly monitors observable variability, capacity, and task performance to identify noise-robust operating points. In addition, we develop a local eigentask analysis that enables computationally efficient feature selection and effective information retrieval. We report evidence of a regime of optimality that is identifiable at small scales and transferable across tasks and larger systems, and we achieve performances competitive with leading classical baselines on representative benchmarks for time-series forecasting and satellite image classification. Together, our results establish a viable and robust framework for large-scale, pre-fault-tolerant quantum machine learning and provide a foundation for extending reservoir-based methods to more expressive architectures and real-world scenarios.

quant-pharXiv:2603.13005•Mar 16, 2026•5 citations

Benchmarking Variational Quantum Regressors Against Classical Linear Models for Solar Power Forecasting

Nandhini Raaj, Benila S

Smart grids form the backbone of modern energy infrastructures. They require accurate power prediction to ensure efficient energy management and maintain grid stability. Recent advancements in quantum computing have provided motivation to explore quantum machine learning models for data driven prediction task. In this study, a quantum computing-based power prediction model is investigated using a Variational Quantum Circuit (VQC) to estimate the AC power output in a solar-based smart grid system. The proposed model is studied using a simulation software created using Qiskit framework and real-world solar dataset is used as input. To enable a holistic comparison, the classical regression models, including Linear Regression, Ridge Regression, Random Forest, and XGBoost, are used as baseline models. Experimental findings indicate that classical models have lower prediction errors and much faster execution times. Conversely, the quantum model can only capture the broad patterns of the data because of the constraints that relate to Noisy Intermediate-Scale Quantum (NISQ) devices such as circuit noise, limited qubit count, and optimization limitations. Furthermore, a multi-criteria evaluation is conducted to compare the classical and quantum approaches based on prediction accuracy, execution time, model complexity, and scalability. The findings provide valuable insights into the feasibility of applying quantum computing techniques to smart grid prediction tasks and highlight both the potential and the limitations of quantum machine learning for real-world energy systems.

International Conference on Recent Advances in Electrical, Electronics, Ubiquitous Communication, and Computational Intelligence (RAEEUCCI)•Apr 22, 2026•0 citations

Exploring Hybrid Quantum-Classical Methods for Practical Time-Series Forecasting

Maksims Dimitrijevs, Mārtiņš Kālis, Ilja Repko

Time-series forecasting is essential for strategic planning and resource allocation, and is important in areas like finance, supply chain management, and energy demand prediction.In such areas, small improvements in forecast quality can lead to significant economic and strategic benefits.It is important to assess the practicality of quantum-enhanced forecasting methods because quantum approaches can complement the strengths of classical models.In this work, we explore two quantum-based approaches for time-series forecasting.The first approach utilizes a parameterized quantum circuit (PQC) model.The second approach employs variational quantum linear regression (VQLS), formulating the forecasting problem as a system of linear equations solved via quantum optimization.A comparative analysis is conducted to assess their practical applicability and performance.In our experiments, the PQC with COBYLA optimization achieved the lowest test error rates, outperforming the classical baselines under the reported settings.VQLS formulation faced stability issues on general synthetic datasets.

arXiv:2412.05615•Baltic Journal of Modern Computing•Jan 1, 2026•2 citations
Introduction Accurate forecasting of solar power generation is essential for grid stability, yet it remains difficult because of the strong volatility and non-linear dynamics of solar resources. Classical machine learning models have improved forecasting performance, but they often fail to capture rapid changes in generation. Methods We propose a hybrid forecasting framework, Convolutional Quantum Extreme Learning Machines (C-QELM), which combines a fixed gate-based quantum feature map with local time-series patches and uses the resulting high-dimensional representations in classical forecasters. To evaluate the approach, we conducted a controlled synthetic benchmark comparing C-QELM with a global Standard QELM, a convolutional purely classical control, and Random Fourier Feature (RFF) maps. We also performed rolling one-step-ahead experiments on three solar datasets, comparing C-QELM with an unaugmented baseline and a dimension-matched classical convolutional random-feature control (CRC). Results In the synthetic benchmark, C-QELM substantially outperformed the global QELM and achieved the lowest mean error among the non-task-informed random-feature methods, although its confidence intervals overlapped with those of CRC and the best RFF configuration. Results on the solar datasets were similarly mixed, indicating that performance depends on both the dataset and the downstream architecture. Discussion Overall, C-QELM does not provide an unconditional performance advantage, but it may reveal a conditional benefit in highly dynamic regimes. Its small register size, low circuit depth, patch-wise reset, and lack of quantum-parameter training make it a promising architecture for future testing on NISQ platforms.
Accurate forecasting of solar power generation plays a vital role in maximising energy output and ensuring grid stability in renewable energy systems. Conventional methods of forecasting can have a hard time modelling the intricate interactions of dynamic environmental factors, like variations in temperature, solar irradiance, and time-of-day effects. This paper proposes a Quantum Machine Learning (QML)-based model that is specifically focused on solar power forecasting and combines three approaches: Quantum Kernel Learning (QKL), K-Nearest Neighbours (KNN) and Linear Regression (LR). Using quantum-enhanced feature mapping and optimised quantum kernels, the model is able to deal with highly non-linear dependencies that are not easily modelled by classical approaches. The proposed approach is named the Hybrid Quantum Classical Model (HQCM) and compared to the classical KNN and LR models in terms of performance parameters. Training and testing were done using data obtained from solar power generation. The simulated results proved that the proposed HQCM can reduce Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) by 57.2% and 55.8%, respectively, with high stability and reliability predictions. The proposed model has achieved a goodR2R^{2}R2of 0.999 with the Quantum K-Nearest Neighbours (QKNN) model and has an absolute error as high as 1.4 MW when the test dataset is recorded. Cumulative Distribution Function (CDF) and residual distribution plots affirm that the quantum model errors are concentrated close to zero, while classical baselines exhibit much larger residual magnitudes. The proposed HQCM has the potential to be helpful in managing the energy grid and making the integration of solar power into smart grids easier. Future work will build on this to real-time optimisation of power grids while considering how other renewable energy sources can be optimised in this way.
Time-series forecasting is vital across sectors such as finance, energy, and environmental science, where accurate future predictions drive informed decision-making. Traditional statistical methods and deep learning models like LSTM and Transformers, while effective, often struggle with high-dimensional data and long-term dependencies. In this paper, we propose a hybrid quantum-classical model that utilizes Parameterized Quantum Circuits (PQCs) to capture complex temporal patterns in multivariate time-series data. Our architecture integrates a classical encoder, a quantum variational processor, and a classical decoder, trained end-to-end using gradient-based optimization. Experiments conducted on benchmark datasets—electricity load, stock prices, and weather forecasting—demonstrate that the hybrid model consistently outperforms classical baselines in terms of Mean Squared Error (MSE), Mean Absolute Error (MAE), and R² score. Robustness analysis confirms the model’s resilience to moderate quantum noise, while an ablation study highlights the performance gain from quantum components. This work showcases the potential of hybrid quantum-classical approaches on Noisy Intermediate-Scale Quantum (NISQ) devices and lays a foundation for scalable, real-world quantum machine learning applications.
Quantum Machine Learning (QML) has emerged as a promising framework for exploring how quantum dynamics may enhance data processing tasks. Here we investigate Quantum Extreme Learning Machines (QELMs), a quantum analogue of classical Extreme Learning Machines in which training is restricted to the output layer. Our architecture combines dimensionality reduction (via PCA or Autoencoders), quantum state encoding, evolution under an XX Hamiltonian, and projective measurement to produce features for a classical single-layer classifier. By analyzing the classification accuracy as a function of evolution time, we observe a sharp transition between low- and high-accuracy regimes, followed by saturation. Remarkably, the saturated performance is comparable to that obtained using Haar-random unitaries that generate maximally complex dynamics, even though the XX model is integrable and local. Our results indicate that this increase in performance correlates with the onset of entanglement, which improves the embedding of classical data in Hilbert space and leads to more separable clusters in measurement probability space. Thus, moderate entanglement can contribute positively to the structure of the data representation, improving learnability without necessarily implying quantum computational advantage. For the image-classification tasks studied here, namely MNIST, Fashion-MNIST, and CIFAR-10, the relevant evolution time is consistent with information exchange over short distances and, within the explored system sizes, does not show evidence of scaling with the full system size. This suggests that QELM performance in this regime relies only on limited entanglement and remains compatible with efficient classical simulation. Our results clarify how local quantum dynamics and moderate quantum correlations are already sufficient to generate useful feature representations for learning.
Reservoir computing leverages rich, non-linear dynamics to process temporal data. Quantum variants promise enhanced expressivity from high-dimensional Hilbert spaces, yet their practical applicability is hindered by hardware noise and concentration effects that can erase input-output distinguishability at large system sizes. In this work, we present and experimentally demonstrate a Quantum Extreme Learning Machine (QELM) tailored to state-of-the-art superconducting platforms, employing up to 124 qubits and circuits with more than 5,000 two-qubit gates on IBM Quantum computers. We introduce a practical multi-objective hyperparameter tuning strategy that jointly monitors observable variability, capacity, and task performance to identify noise-robust operating points. In addition, we develop a local eigentask analysis that enables computationally efficient feature selection and effective information retrieval. We report evidence of a regime of optimality that is identifiable at small scales and transferable across tasks and larger systems, and we achieve performances competitive with leading classical baselines on representative benchmarks for time-series forecasting and satellite image classification. Together, our results establish a viable and robust framework for large-scale, pre-fault-tolerant quantum machine learning and provide a foundation for extending reservoir-based methods to more expressive architectures and real-world scenarios.
Smart grids form the backbone of modern energy infrastructures. They require accurate power prediction to ensure efficient energy management and maintain grid stability. Recent advancements in quantum computing have provided motivation to explore quantum machine learning models for data driven prediction task. In this study, a quantum computing-based power prediction model is investigated using a Variational Quantum Circuit (VQC) to estimate the AC power output in a solar-based smart grid system. The proposed model is studied using a simulation software created using Qiskit framework and real-world solar dataset is used as input. To enable a holistic comparison, the classical regression models, including Linear Regression, Ridge Regression, Random Forest, and XGBoost, are used as baseline models. Experimental findings indicate that classical models have lower prediction errors and much faster execution times. Conversely, the quantum model can only capture the broad patterns of the data because of the constraints that relate to Noisy Intermediate-Scale Quantum (NISQ) devices such as circuit noise, limited qubit count, and optimization limitations. Furthermore, a multi-criteria evaluation is conducted to compare the classical and quantum approaches based on prediction accuracy, execution time, model complexity, and scalability. The findings provide valuable insights into the feasibility of applying quantum computing techniques to smart grid prediction tasks and highlight both the potential and the limitations of quantum machine learning for real-world energy systems.
Time-series forecasting is essential for strategic planning and resource allocation, and is important in areas like finance, supply chain management, and energy demand prediction.In such areas, small improvements in forecast quality can lead to significant economic and strategic benefits.It is important to assess the practicality of quantum-enhanced forecasting methods because quantum approaches can complement the strengths of classical models.In this work, we explore two quantum-based approaches for time-series forecasting.The first approach utilizes a parameterized quantum circuit (PQC) model.The second approach employs variational quantum linear regression (VQLS), formulating the forecasting problem as a system of linear equations solved via quantum optimization.A comparative analysis is conducted to assess their practical applicability and performance.In our experiments, the PQC with COBYLA optimization achieved the lowest test error rates, outperforming the classical baselines under the reported settings.VQLS formulation faced stability issues on general synthetic datasets.