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

IEEE Access

Solar Irradiance Forecasting Using a Hybrid Quantum Neural Network: A Comparison on GPU-Based Workflow Development Platforms

Ying‐Yi Hong, Dylan Josh Domingo Lopez, Yun-Yuan Wang

Published Jan 1, 2024• 27 citations
Publisher

Abstract

Modern renewable power operations can be enhanced by integrating deep neural networks, particularly for forecasting solar irradiance. Recent advancements in quantum computing have shown potential improvements in classical deep neural networks. However, current challenges with quantum hardware, such as susceptibility to noise and decoherence, pose risks to its practicality. Hybrid quantum neural networks (HQNNs) are found to mitigate these issues, especially when integrated with graphics processing unit (GPU)-based pipelines. This paper presents a comparative study of different software platforms for developing HQNNs, using multi-location very short-term solar irradiance forecasting as an example. A classical benchmark model is initially designed based on statistical analysis of a 10-minute resolution solar irradiance dataset, with its parameters further optimized using Bayesian Optimization. The experimental design of this paper includes a loss comparison between classical neural networks and HQNNs across different seasons and a performance comparison between Pennylane, Torchquantum, and CUDA Quantum (CUDA-Q) as HQNN development platforms. Experimental results show that HQNNs achieve up to a 92.30% improvement in testing loss compared to classical neural networks. Regarding HQNN development platforms, Pennylane shows an 81.54% testing loss reduction from classical models, Torchquantum shows a 90.34% improvement, and CUDA-Q shows a 92.30% improvement in testing loss. Implementing hardware acceleration libraries for GPU-based state vector simulation demonstrates an approximate 275% speedup in average latency per epoch, a 218% speedup in inference time, and a 10.20% improvement in testing loss compared to CPU-based simulations. CUDA-Q achieves a training time 2.7 times shorter and an inference time 2.9 times shorter compared to Pennylane, while it is 32.3 times faster in training and 31 times faster in inference compared to Torchquantum.

DOI
10.1109/access.2024.3472053

Related papers

Improved quantum long short-term memory with successive variational mode decomposition for solar irradiance prediction

Snigdha Sarangi, P.K. Dash, Ranjeeta Bisoi

Short-term solar irradiance prediction is critical for efficient power scheduling, the design of solar-integrated energy management systems, and maintaining grid stability. However, the inherent variability of solar energy introduces fluctuations in power output, and posing challenges for reliable grid operation. Leveraging recent advancements in quantum computing, the current work proposes a hybrid quantum-driven neural network for solar irradiance prediction, combining successive variational mode decomposition (SVMD) with an improved quantum long short-term memory (IQLSTM) network. To enhance circuit expressibility, a robust variational quantum circuit (RVQC) incorporating controlled-X (CRX) gates is embedded within the quantum long short-term memory (QLSTM) architecture. In the RVQC, a circular topology with two-way connections is employed to optimize qubit interconnections and strengthen entanglement. The proposed model is evaluated using two solar irradiance datasets of Hyderabad (2019) and Bengaluru (2019, 2018), obtained from the National Renewable Energy Laboratory (NREL). To assess the effectiveness of the proposed circuit, evaluations are carried out under varying weather conditions and across multiple forecasting horizons, and results are compared against classical LSTM and other quantum-based networks like quantum gate recurrent units (QGRUs), quantum support vector machine (QSVM) and QLSTM. Furthermore, the model is evaluated under noisy quantum environments to approximate realistic Noisy Intermediate-Scale Quantum (NISQ) hardware conditions, enabling the assessment of its robustness under practical quantum computing constraints. Experimental results demonstrate that the proposed model achieves superior predictive accuracy and robustness compared with existing classical and quantum-based approaches.

Scientific Reports•May 19, 2026•0 citations

Implementing a Hybrid Quantum Neural Network for Wind Speed Forecasting: Insights from Quantum Simulator Experiences

Ying‐Yi Hong, Jay Bhie D. Santos

The intermittent nature of wind speed poses challenges for its widespread utilization as an electrical power generation source. As the integration of wind energy into the power system increases, accurate wind speed forecasting becomes crucial. The reliable scheduling of wind power generation heavily relies on precise wind speed forecasts. This paper presents an extended work that focuses on a hybrid model for 24 h ahead wind speed forecasting. The proposed model combines residual Long Short-Term Memory (LSTM) and a quantum neural network that is studied by a quantum simulator, leveraging the support of NVIDIA Compute Unified Device Architecture (CUDA). To ensure the desired accuracy, a comparative analysis is conducted, examining the qubit count and quantum circuit depth of the proposed model. The execution time required for the model is significantly reduced when the GPU incorporates CUDA, accounting for only 8.29% of the time required by a classical CPU. In addition, different quantum embedding layers with various entangler layers in the quantum neural network are explored. The simulation results utilizing an offshore wind farm dataset demonstrate that the proper number of qubits and embedding layer can achieve favorable 24 h ahead wind speed forecasts.

Energies•Apr 1, 2025•2 citations

GPU-Based Hybrid Quantum Computing for EV Charging Station Decision Planning–A Case Study

Dylan Josh Domingo Lopez, Ying-Yi Hong, Mark Angelo Purio +4 more

Electric vehicle charging station (EVCS) planning increasingly relies on data analytics pipelines. Yet current approaches remain purely classical and do not exploit emerging hybrid quantum methods for coupled forecasting and optimization. This paper proposes a graphics processing unit (GPU)-based hybrid quantum computing pipeline that integrates quantum-enhanced forecasting and decision models into an end-to-end planning workflow. For forecasting, a hybrid quantum neural network augments a long short-term memory network forecaster with a quantum head, while GPU-parallelized parameter-shift gradients and batched inference reduce training time by up to 76.9% relative to a non-parallel baseline without sacrificing test-set accuracy. For decision optimization, a station decommissioning problem is formulated as a mixed quadratic integer program, converted to a scaled quadratic unconstrained binary optimization model, and then solved via both classical annealing and quantum approximate optimization algorithm using a toolchain spanning CVXPY, autoQUBO/OpenJij, Qamomile, Qiskit, and CUDA Q/X, with scaled model runs matching the classical optimal profit and configuration while benefiting from GPU-accelerated simulation. At the 15-qubit scale studied under noiseless simulation, hybrid quantum components match but do not exceed classical baselines; the evidence supports their use as weighted heads or optimization backends co-designed with GPU parallelization and classical solvers, with quantum advantage and noise robustness deferred to larger-scale and QPU-based future work.

IEEE International Conference on Quantum Control, Computing and Learning (qCCL)•Jun 1, 2026•0 citations

Comparative Analysis of Hybrid Quantum-Classical and Artificial Neural Network for Sustainable Cyber-Physical Power System Anomaly Detection

Lloyd Aldrin T. PORNOBI

Abstract. Smart grid infrastructures are becoming increasingly vulnerable to a range of physical and cyber-physical faults that can result in outages, equipment damage, and safety hazards. This paper presents a comparative evaluation of a hybrid quantum-classical neural network (HQNN) against a classical artificial neural network (ANN) for binary fault detection in a simulated smart-grid environment. Both models are trained and tested on a 4,965-sample dataset derived from a power-system simulator jointly developed by Mississippi State University and Oak Ridge National Laboratory. We encode input features into quantum states using Hadamard and RY rotations, employ entangling layers in PennyLane, and fuse the resulting quantum expectation values with classical fully connected layers. The ANN baseline, in contrast, utilizes two ReLU-activated hidden layers and dropout. Performance is measured in terms of precision, recall, F1-score, ROC AUC, and CPU-only inference latency, with 5-fold cross-validation implemented. Results show that the HQNN got a higher ROC AUC (0.7565 vs. 0.6763), higher precision (80% vs. 69%), recall (61% vs. 54%), and F1-Score (65% vs. 58%), while the ANN achieves a lower average inference time that runs almost 65x faster per sample. These findings demonstrate that hybrid quantum-classical models can enhance discriminative power for sustainable, low-power smart-grid fault detection, even in resource-limited settings, at the cost of slightly increased inference time.

Materials research proceedings•Jan 1, 2026•0 citations

Quantum-Classical Neural Networks: A Hybrid Approach to MNIST Classification

Tur Mohammad Mubtasim, Md. Siam Uddin Molla Antor, Sheikh Shifat

The combination of the principles of quantum computing and classical deep learning is a potentially fruitful direction to develop the field of artificial intelligence, but the benchmarking of such hybrid architectures against standard computer vision problems is an under-explored area. The paper explores the behavior of Hybrid Quantum-Classical Neural Network (HQNN) in the classification of images into multiple classes using the MNIST database. Our architecture is an HQNN implementing an 8-qubit and 2-layer variational circuit, and it is parameterized by a classical multi-layer perceptron (MLP). To make the evaluation process rigorous and fair, the hybrid was benchmarked on a purely classical MLP, which had almost the same parameter footprint ( 105,000 parameters). Results over 50 training epochs show that the classical MLP reached its highest test accuracy of 96.50% and Macro F1 of 0.96. The HQNN was able to converge and learn the data of high dimensions delivering a competitive but slightly lower maximum accuracy of 94.35% and Macro F1 of 0.94. Moreover, there was a high computational cost on the simulation of the quantum layers, which increased the training time by 233 times the classical baseline. These results suggest that while the quantum layer is expressive and capable of learning complex representations, the current state of quantum simulation on classical hardware is a significant bottleneck. Which can be reduced by QPU/NISQ hardware, lighter quantum circuits and better dta encoding.

International Conference on Power, Electronics, Communications, Computing, and Intelligent Infrastructure (PECCII)•Jun 1, 2026•0 citations

Distributed hybrid quantum-classical performance prediction for hyperparameter optimization

Eric Wulff, Juan Pablo García Amboage, Marcel Aach +9 more

Abstract Hyperparameter optimization (HPO) of neural networks is a computationally expensive procedure, which requires a large number of different model configurations to be trained. To reduce such costs, this work presents a distributed, hybrid workflow, that runs the training of the neural networks on multiple graphics processing units (GPUs) on a classical supercomputer, while predicting the configurations’ performance with quantum-trained support vector regression (QT-SVR) on a quantum annealer (QA). The workflow is shown to run on up to 50 GPUs and a QA at the same time, completely automating the communication between the classical and the quantum systems. The approach is evaluated extensively on several benchmarking datasets from the computer vision (CV), high-energy physics (HEP), and natural language processing (NLP) domains. Empirical results show that resource costs for performing HPO can be reduced by up to 9% when using the hybrid workflow with performance prediction, compared to using a plain HPO algorithm without performance prediction. Additionally, the workflow obtains similar and in some cases even better accuracy of the final hyperparameter configuration, when combining multiple heuristically obtained predictions from the QA, compared to using just a single classically obtained prediction. The results highlight the potential of hybrid quantum-classical machine learning algorithms. The workflow code is made available open-source to foster adoption in the community.

Quantum Machine Intelligence•Sep 14, 2024•8 citations
Modern renewable power operations can be enhanced by integrating deep neural networks, particularly for forecasting solar irradiance. Recent advancements in quantum computing have shown potential improvements in classical deep neural networks. However, current challenges with quantum hardware, such as susceptibility to noise and decoherence, pose risks to its practicality. Hybrid quantum neural networks (HQNNs) are found to mitigate these issues, especially when integrated with graphics processing unit (GPU)-based pipelines. This paper presents a comparative study of different software platforms for developing HQNNs, using multi-location very short-term solar irradiance forecasting as an example. A classical benchmark model is initially designed based on statistical analysis of a 10-minute resolution solar irradiance dataset, with its parameters further optimized using Bayesian Optimization. The experimental design of this paper includes a loss comparison between classical neural networks and HQNNs across different seasons and a performance comparison between Pennylane, Torchquantum, and CUDA Quantum (CUDA-Q) as HQNN development platforms. Experimental results show that HQNNs achieve up to a 92.30% improvement in testing loss compared to classical neural networks. Regarding HQNN development platforms, Pennylane shows an 81.54% testing loss reduction from classical models, Torchquantum shows a 90.34% improvement, and CUDA-Q shows a 92.30% improvement in testing loss. Implementing hardware acceleration libraries for GPU-based state vector simulation demonstrates an approximate 275% speedup in average latency per epoch, a 218% speedup in inference time, and a 10.20% improvement in testing loss compared to CPU-based simulations. CUDA-Q achieves a training time 2.7 times shorter and an inference time 2.9 times shorter compared to Pennylane, while it is 32.3 times faster in training and 31 times faster in inference compared to Torchquantum.
Short-term solar irradiance prediction is critical for efficient power scheduling, the design of solar-integrated energy management systems, and maintaining grid stability. However, the inherent variability of solar energy introduces fluctuations in power output, and posing challenges for reliable grid operation. Leveraging recent advancements in quantum computing, the current work proposes a hybrid quantum-driven neural network for solar irradiance prediction, combining successive variational mode decomposition (SVMD) with an improved quantum long short-term memory (IQLSTM) network. To enhance circuit expressibility, a robust variational quantum circuit (RVQC) incorporating controlled-X (CRX) gates is embedded within the quantum long short-term memory (QLSTM) architecture. In the RVQC, a circular topology with two-way connections is employed to optimize qubit interconnections and strengthen entanglement. The proposed model is evaluated using two solar irradiance datasets of Hyderabad (2019) and Bengaluru (2019, 2018), obtained from the National Renewable Energy Laboratory (NREL). To assess the effectiveness of the proposed circuit, evaluations are carried out under varying weather conditions and across multiple forecasting horizons, and results are compared against classical LSTM and other quantum-based networks like quantum gate recurrent units (QGRUs), quantum support vector machine (QSVM) and QLSTM. Furthermore, the model is evaluated under noisy quantum environments to approximate realistic Noisy Intermediate-Scale Quantum (NISQ) hardware conditions, enabling the assessment of its robustness under practical quantum computing constraints. Experimental results demonstrate that the proposed model achieves superior predictive accuracy and robustness compared with existing classical and quantum-based approaches.
The intermittent nature of wind speed poses challenges for its widespread utilization as an electrical power generation source. As the integration of wind energy into the power system increases, accurate wind speed forecasting becomes crucial. The reliable scheduling of wind power generation heavily relies on precise wind speed forecasts. This paper presents an extended work that focuses on a hybrid model for 24 h ahead wind speed forecasting. The proposed model combines residual Long Short-Term Memory (LSTM) and a quantum neural network that is studied by a quantum simulator, leveraging the support of NVIDIA Compute Unified Device Architecture (CUDA). To ensure the desired accuracy, a comparative analysis is conducted, examining the qubit count and quantum circuit depth of the proposed model. The execution time required for the model is significantly reduced when the GPU incorporates CUDA, accounting for only 8.29% of the time required by a classical CPU. In addition, different quantum embedding layers with various entangler layers in the quantum neural network are explored. The simulation results utilizing an offshore wind farm dataset demonstrate that the proper number of qubits and embedding layer can achieve favorable 24 h ahead wind speed forecasts.
Electric vehicle charging station (EVCS) planning increasingly relies on data analytics pipelines. Yet current approaches remain purely classical and do not exploit emerging hybrid quantum methods for coupled forecasting and optimization. This paper proposes a graphics processing unit (GPU)-based hybrid quantum computing pipeline that integrates quantum-enhanced forecasting and decision models into an end-to-end planning workflow. For forecasting, a hybrid quantum neural network augments a long short-term memory network forecaster with a quantum head, while GPU-parallelized parameter-shift gradients and batched inference reduce training time by up to 76.9% relative to a non-parallel baseline without sacrificing test-set accuracy. For decision optimization, a station decommissioning problem is formulated as a mixed quadratic integer program, converted to a scaled quadratic unconstrained binary optimization model, and then solved via both classical annealing and quantum approximate optimization algorithm using a toolchain spanning CVXPY, autoQUBO/OpenJij, Qamomile, Qiskit, and CUDA Q/X, with scaled model runs matching the classical optimal profit and configuration while benefiting from GPU-accelerated simulation. At the 15-qubit scale studied under noiseless simulation, hybrid quantum components match but do not exceed classical baselines; the evidence supports their use as weighted heads or optimization backends co-designed with GPU parallelization and classical solvers, with quantum advantage and noise robustness deferred to larger-scale and QPU-based future work.
Abstract. Smart grid infrastructures are becoming increasingly vulnerable to a range of physical and cyber-physical faults that can result in outages, equipment damage, and safety hazards. This paper presents a comparative evaluation of a hybrid quantum-classical neural network (HQNN) against a classical artificial neural network (ANN) for binary fault detection in a simulated smart-grid environment. Both models are trained and tested on a 4,965-sample dataset derived from a power-system simulator jointly developed by Mississippi State University and Oak Ridge National Laboratory. We encode input features into quantum states using Hadamard and RY rotations, employ entangling layers in PennyLane, and fuse the resulting quantum expectation values with classical fully connected layers. The ANN baseline, in contrast, utilizes two ReLU-activated hidden layers and dropout. Performance is measured in terms of precision, recall, F1-score, ROC AUC, and CPU-only inference latency, with 5-fold cross-validation implemented. Results show that the HQNN got a higher ROC AUC (0.7565 vs. 0.6763), higher precision (80% vs. 69%), recall (61% vs. 54%), and F1-Score (65% vs. 58%), while the ANN achieves a lower average inference time that runs almost 65x faster per sample. These findings demonstrate that hybrid quantum-classical models can enhance discriminative power for sustainable, low-power smart-grid fault detection, even in resource-limited settings, at the cost of slightly increased inference time.
The combination of the principles of quantum computing and classical deep learning is a potentially fruitful direction to develop the field of artificial intelligence, but the benchmarking of such hybrid architectures against standard computer vision problems is an under-explored area. The paper explores the behavior of Hybrid Quantum-Classical Neural Network (HQNN) in the classification of images into multiple classes using the MNIST database. Our architecture is an HQNN implementing an 8-qubit and 2-layer variational circuit, and it is parameterized by a classical multi-layer perceptron (MLP). To make the evaluation process rigorous and fair, the hybrid was benchmarked on a purely classical MLP, which had almost the same parameter footprint ( 105,000 parameters). Results over 50 training epochs show that the classical MLP reached its highest test accuracy of 96.50% and Macro F1 of 0.96. The HQNN was able to converge and learn the data of high dimensions delivering a competitive but slightly lower maximum accuracy of 94.35% and Macro F1 of 0.94. Moreover, there was a high computational cost on the simulation of the quantum layers, which increased the training time by 233 times the classical baseline. These results suggest that while the quantum layer is expressive and capable of learning complex representations, the current state of quantum simulation on classical hardware is a significant bottleneck. Which can be reduced by QPU/NISQ hardware, lighter quantum circuits and better dta encoding.
Abstract Hyperparameter optimization (HPO) of neural networks is a computationally expensive procedure, which requires a large number of different model configurations to be trained. To reduce such costs, this work presents a distributed, hybrid workflow, that runs the training of the neural networks on multiple graphics processing units (GPUs) on a classical supercomputer, while predicting the configurations’ performance with quantum-trained support vector regression (QT-SVR) on a quantum annealer (QA). The workflow is shown to run on up to 50 GPUs and a QA at the same time, completely automating the communication between the classical and the quantum systems. The approach is evaluated extensively on several benchmarking datasets from the computer vision (CV), high-energy physics (HEP), and natural language processing (NLP) domains. Empirical results show that resource costs for performing HPO can be reduced by up to 9% when using the hybrid workflow with performance prediction, compared to using a plain HPO algorithm without performance prediction. Additionally, the workflow obtains similar and in some cases even better accuracy of the final hyperparameter configuration, when combining multiple heuristically obtained predictions from the QA, compared to using just a single classically obtained prediction. The results highlight the potential of hybrid quantum-classical machine learning algorithms. The workflow code is made available open-source to foster adoption in the community.