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

Quantum machine learning performance evaluation for air pollution assessment with benchmarking and comparative analysis using multi-metric scoring and quantum circuit expressibility

Jagadish Kumar Mogaraju

Published Jul 9, 2026• 0 citations
Publisher

Abstract

Abstract Our study proposes the MDQ-3L framework, a multi-metric, quantum-aware hierarchy for benchmarking classical machine learning, deep learning, and quantum machine learning models in air quality prediction. PM2.5, PM10, NO, NO2, NOx, NH3, CO, SO2, O3, Benzene, Toluene, and Xylene were considered predictors of AQI. Seventeen ML regressors, nine DL models, and seven QML circuits were evaluated using MAE, MSE, RMSE, R², RMSLE, MAPE, and QML-specific expressibility (0.739-2.0) and entangling power (0-0.398). Aggregated via min-max normalization and weights, the framework penalizes negative R² and high structural costs, yielding CatBoost and DBN as top performers, while QML trailed by 25–40% in composite scores due to generalization gaps and barren plateau risks. MDQ-3L aligns circuits to targets, reducing QCNN/TTN penalties by ~ 30% relative to non-benchmarked rankings. Our approach exposes QML limitations under NISQ constraints (4–6 qubits, shallow depth), supporting low-entanglement ansätze like SELR/HEA for 10–15% potential gains, and provides an interpretable tool for hybrid environmental modeling, advancing Quantum Machine Intelligence benchmarks.

DOI
10.21203/rs.3.rs-10290235/v1

Related papers

Comparative Analysis of Neural Network and Quantum Neural Network for Enhanced Air Quality Forecasting

Noha A. Sobhi, Doaa A. Shoieb, Sherin Moustafa Youssef

Air pollution forecasting is a critical challenge in environmental science. Nitrogen Dioxide (NO2) and Carbon Monoxide (CO) are two of the most harmful gases, contributing significantly to respiratory illnesses and environmental degradation. Accurate predictions of NO2and CO levels help mitigate the adverse effects of pollutants on public health. A hybrid quantum-classical machine learning model is developed, integrating a feature map, variational ansatz, and quantum regressor optimized with multiple optimizers. The results are compared with classical machine learning model, such as Multi-Layer Perceptron (MLP), demonstrating the potential of quantum computing in predictive modeling. Experimental results using the air quality dataset reveal that the proposed model achieves superior performance compared to traditional approaches, as measured by MSE, RMSE, MAE, and R2. The findings suggest that quantum-enhanced neural networks provide a promising avenue for real-time air pollution forecasting, enabling proactive pollution control measures.

International Conference on Machine Intelligence and Smart Innovation (ICMISI)•May 10, 2025•1 citation

Hybrid quantum-classical machine learning with SABHQ for groundwater quality assessment

Jagadish Kumar Mogaraju

Quantum Machine Learning is a growing field with the potential to unlock several opportunities in groundwater management. A novel approach integrating advanced quantum machine learning (QML) models with classical machine learning (ML) algorithms for robust groundwater quality assessment, using the SABHQ (Systematically Averaged Battery Hybrid Quantum) method, is proposed. We attempted to reduce noise and enhance generalization by stimulating composite sampling across both ML and QML pipelines. Accuracy, precision, recall, and F 1 scores were calculated across all models, using normalization and aggregation strategies to achieve relatively unbiased benchmarking. We observed that the dominant predictors, such as EC, Cl, HCO 3 , SO 4 , Na, TH, Latitude, and Longitude, reflect both spatial variability and anthropogenic influences on groundwater quality. The systematic averaging method yielded noticeable improvements for QML models, with QCNN achieving accuracy and F 1 scores of up to 0.95 and 0.98, respectively, outperforming classical machine learning models. We report that QCNN metrics improved due to aggregation, noise reduction, and possibly quantum entanglement effects, leading to better generalization, whereas VQC showed gradual but minor improvements. We believe our work can be extended to other environmental domains, not only to refine the method we proposed but also to enhance its adaptability.

Discover Quantum Science•Mar 30, 2026•0 citations

Q-SCALE: Quantum Computing-Based Sensor Calibration for Advanced Learning and Efficiency

Lorenzo Bergadano, Andrea Ceschini, Pietro Chiavassa +4 more

In a world burdened by air pollution, the integration of state-of-the-art sensor calibration techniques utilizing Quantum Computing (QC) and Machine Learning (ML) holds promise for enhancing the accuracy and efficiency of air quality monitoring systems in smart cities. This article investigates the process of calibrating inexpensive optical fine-dust sensors through advanced methodologies such as Deep Learning (DL) and Quantum Machine Learning (QML). The objective of the project is to compare four sophisticated algorithms from both the classical and quantum realms to discern their disparities and explore possible alternative approaches to improve the precision and dependability of particulate matter measurements in urban air quality surveillance. Classical Feed-Forward Neural Networks (FFNN) and Long Short-Term Memory (LSTM) models are evaluated against their quantum counterparts: Variational Quantum Regressors (VQR) and Quantum LSTM (QLSTM) circuits. Through meticulous testing, including hyperparameter optimization and cross-validation, the study assesses the potential of quantum models to refine calibration performance. Our analysis shows that: the FFNN model achieved superior calibration accuracy on the test set compared to the VQR model in terms of lower L1 loss function (2.92 vs 4.81); the QLSTM slightly outperformed the LSTM model (loss on the test set: 2.70 vs 2.77), despite using fewer trainable weights (66 vs 482).

arXiv:2410.02998•IEEE International Conference on Quantum Computing and Engineering (QCE)•Sep 15, 2024•3 citations

Benchmarking Quantum Machine Learning Models Against Classical Deep Learning

Elena Silva, Amelia Horvath

Quantum machine learning (QML) benchmarks frequently compare quantum models against weak classical baselines -- the same methodological gap identified for quantum optimisation in the QCCB study (Popescu and Petrov, 2024). This paper presents the QML-DL Benchmark (QMLDB), a rigorous comparative evaluation of six QML model families against nine state-of-the-art classical deep learning baselines (MLP, CNN, LSTM, ResNet, Transformer, TabNet, XGBoost, LightGBM, CatBoost) across ten benchmark datasets spanning six task domains. QMLDB controls for model capacity, training budget, and data size. QML achieves top performance on three of ten datasets -- all three with strong quantum-process proximity. On the remaining seven, classical deep learning dominates, with Transformer and ResNet consistently outperforming all QML models by 4.8-18.4pp. QMLDB establishes that QML advantage is narrower than commonly claimed, confined to datasets with genuine quantum structure, and proposes a quantum-process proximity score (QPPS) to predict whether a new dataset will benefit from QML before hardware experiments.

Quantum Frontiers Journal•Mar 28, 2025•0 citations

Hybrid Quantum-Classical Machine Learning Models: Design, Implementation, And Performance Evaluation On NISQ Devices

P. Sunil, G. Swapna

Quantum machine learning has emerged as a promising approach to enhance computational efficiency by leveraging the principles of quantum computing. However, the limitations of current Noisy Intermediate-Scale Quantum (NISQ) devices, such as noise, limited qubit availability, and circuit depth constraints, restrict the implementation of fully quantum models. To address these challenges, this study focuses on the design, implementation, and performance evaluation of hybrid quantum-classical machine learning (HQML) models. The proposed approach integrates parameterized quantum circuits with classical optimization techniques to enable efficient learning within NISQ environments. The study employs standard benchmark datasets, including Iris, Breast Cancer, and MNIST, to evaluate the performance of the hybrid model. The results indicate that the HQML model achieves competitive accuracy on small and medium-sized datasets while maintaining balanced precision, recall, and F1-score. However, performance declines for complex datasets due to hardware limitations and noise effects. Additionally, the hybrid model demonstrates a lower number of parameters compared to classical deep learning models but requires higher training time due to iterative quantum-classical optimization. The findings highlight that hybrid quantum-classical models provide a practical and scalable approach for utilizing quantum computing in the current technological landscape. Although challenges related to noise, scalability, and computational overhead persist, advancements in quantum hardware and algorithm design are expected to improve performance. This study contributes to the growing field of quantum machine learning by providing a systematic framework for evaluating hybrid models on NISQ devices and identifying key areas for future research.

•Jan 1, 2026•0 citations

Evaluating Classical, Hybrid, and NISQ-Era Quantum Regression Models for Atmospheric Pollutant Concentration Prediction

P G Athira, M P Deepika

: Quantum physics-based quantum computing has gained attention as a new computing paradigm. As the quantum technology is advancing the Quantum Machine Learning (QML) is emerging as a new research area that combines quantum computing and machine learning. As classical machine learning continues to advance rapidly, by exploring new computing paradigms quantum machine learning is also progressing by taking a new parallel direction. In this paper, we implement and compare classical machine learning model, hybrid quantum-classical model, and fully quantum regression models for predicting Atmospheric Pollutant Concentration (APC). Multiple Linear Regression and Support Vector Regression are used as standard classical methods. Two hybrid models- Quantum Kernel Regression and Hybrid Quantum–Classical Regression, and a fully quantum model -Variational Quantum Regression are examined to understand how quantum techniques perform on today’s noisy quantum computers, known as Noisy Intermediate-Scale Quantum (NISQ) devices. The implementation and testing are conducted using a real time atmospheric pollutant concentration dataset containing data from different states of India, which is collected from Open Government Data (OGD) platform. The quantum models are implemented using quantum circuits and simulated using IBM Qiskit. The studies show that in this NISQ era- mainly limited qubits which are noisy and limited depth circuits, the classical regression models give more accurate predictions than quantum models. However, the quantum and hybrid approaches show comparable training times and provide useful insights into how these methods may scale in the future. Overall, this study highlights both the potential and the current limitations of quantum regression techniques when applied to real-world environmental prediction tasks in the NISQ era.

International Conference on Advances in Intelligent Computing and Applications (AICAPS)•Feb 11, 2026•0 citations
Abstract Our study proposes the MDQ-3L framework, a multi-metric, quantum-aware hierarchy for benchmarking classical machine learning, deep learning, and quantum machine learning models in air quality prediction. PM2.5, PM10, NO, NO2, NOx, NH3, CO, SO2, O3, Benzene, Toluene, and Xylene were considered predictors of AQI. Seventeen ML regressors, nine DL models, and seven QML circuits were evaluated using MAE, MSE, RMSE, R², RMSLE, MAPE, and QML-specific expressibility (0.739-2.0) and entangling power (0-0.398). Aggregated via min-max normalization and weights, the framework penalizes negative R² and high structural costs, yielding CatBoost and DBN as top performers, while QML trailed by 25–40% in composite scores due to generalization gaps and barren plateau risks. MDQ-3L aligns circuits to targets, reducing QCNN/TTN penalties by ~ 30% relative to non-benchmarked rankings. Our approach exposes QML limitations under NISQ constraints (4–6 qubits, shallow depth), supporting low-entanglement ansätze like SELR/HEA for 10–15% potential gains, and provides an interpretable tool for hybrid environmental modeling, advancing Quantum Machine Intelligence benchmarks.
Air pollution forecasting is a critical challenge in environmental science. Nitrogen Dioxide (NO2) and Carbon Monoxide (CO) are two of the most harmful gases, contributing significantly to respiratory illnesses and environmental degradation. Accurate predictions of NO2and CO levels help mitigate the adverse effects of pollutants on public health. A hybrid quantum-classical machine learning model is developed, integrating a feature map, variational ansatz, and quantum regressor optimized with multiple optimizers. The results are compared with classical machine learning model, such as Multi-Layer Perceptron (MLP), demonstrating the potential of quantum computing in predictive modeling. Experimental results using the air quality dataset reveal that the proposed model achieves superior performance compared to traditional approaches, as measured by MSE, RMSE, MAE, and R2. The findings suggest that quantum-enhanced neural networks provide a promising avenue for real-time air pollution forecasting, enabling proactive pollution control measures.
Quantum Machine Learning is a growing field with the potential to unlock several opportunities in groundwater management. A novel approach integrating advanced quantum machine learning (QML) models with classical machine learning (ML) algorithms for robust groundwater quality assessment, using the SABHQ (Systematically Averaged Battery Hybrid Quantum) method, is proposed. We attempted to reduce noise and enhance generalization by stimulating composite sampling across both ML and QML pipelines. Accuracy, precision, recall, and F 1 scores were calculated across all models, using normalization and aggregation strategies to achieve relatively unbiased benchmarking. We observed that the dominant predictors, such as EC, Cl, HCO 3 , SO 4 , Na, TH, Latitude, and Longitude, reflect both spatial variability and anthropogenic influences on groundwater quality. The systematic averaging method yielded noticeable improvements for QML models, with QCNN achieving accuracy and F 1 scores of up to 0.95 and 0.98, respectively, outperforming classical machine learning models. We report that QCNN metrics improved due to aggregation, noise reduction, and possibly quantum entanglement effects, leading to better generalization, whereas VQC showed gradual but minor improvements. We believe our work can be extended to other environmental domains, not only to refine the method we proposed but also to enhance its adaptability.
In a world burdened by air pollution, the integration of state-of-the-art sensor calibration techniques utilizing Quantum Computing (QC) and Machine Learning (ML) holds promise for enhancing the accuracy and efficiency of air quality monitoring systems in smart cities. This article investigates the process of calibrating inexpensive optical fine-dust sensors through advanced methodologies such as Deep Learning (DL) and Quantum Machine Learning (QML). The objective of the project is to compare four sophisticated algorithms from both the classical and quantum realms to discern their disparities and explore possible alternative approaches to improve the precision and dependability of particulate matter measurements in urban air quality surveillance. Classical Feed-Forward Neural Networks (FFNN) and Long Short-Term Memory (LSTM) models are evaluated against their quantum counterparts: Variational Quantum Regressors (VQR) and Quantum LSTM (QLSTM) circuits. Through meticulous testing, including hyperparameter optimization and cross-validation, the study assesses the potential of quantum models to refine calibration performance. Our analysis shows that: the FFNN model achieved superior calibration accuracy on the test set compared to the VQR model in terms of lower L1 loss function (2.92 vs 4.81); the QLSTM slightly outperformed the LSTM model (loss on the test set: 2.70 vs 2.77), despite using fewer trainable weights (66 vs 482).
Quantum machine learning (QML) benchmarks frequently compare quantum models against weak classical baselines -- the same methodological gap identified for quantum optimisation in the QCCB study (Popescu and Petrov, 2024). This paper presents the QML-DL Benchmark (QMLDB), a rigorous comparative evaluation of six QML model families against nine state-of-the-art classical deep learning baselines (MLP, CNN, LSTM, ResNet, Transformer, TabNet, XGBoost, LightGBM, CatBoost) across ten benchmark datasets spanning six task domains. QMLDB controls for model capacity, training budget, and data size. QML achieves top performance on three of ten datasets -- all three with strong quantum-process proximity. On the remaining seven, classical deep learning dominates, with Transformer and ResNet consistently outperforming all QML models by 4.8-18.4pp. QMLDB establishes that QML advantage is narrower than commonly claimed, confined to datasets with genuine quantum structure, and proposes a quantum-process proximity score (QPPS) to predict whether a new dataset will benefit from QML before hardware experiments.
Quantum machine learning has emerged as a promising approach to enhance computational efficiency by leveraging the principles of quantum computing. However, the limitations of current Noisy Intermediate-Scale Quantum (NISQ) devices, such as noise, limited qubit availability, and circuit depth constraints, restrict the implementation of fully quantum models. To address these challenges, this study focuses on the design, implementation, and performance evaluation of hybrid quantum-classical machine learning (HQML) models. The proposed approach integrates parameterized quantum circuits with classical optimization techniques to enable efficient learning within NISQ environments. The study employs standard benchmark datasets, including Iris, Breast Cancer, and MNIST, to evaluate the performance of the hybrid model. The results indicate that the HQML model achieves competitive accuracy on small and medium-sized datasets while maintaining balanced precision, recall, and F1-score. However, performance declines for complex datasets due to hardware limitations and noise effects. Additionally, the hybrid model demonstrates a lower number of parameters compared to classical deep learning models but requires higher training time due to iterative quantum-classical optimization. The findings highlight that hybrid quantum-classical models provide a practical and scalable approach for utilizing quantum computing in the current technological landscape. Although challenges related to noise, scalability, and computational overhead persist, advancements in quantum hardware and algorithm design are expected to improve performance. This study contributes to the growing field of quantum machine learning by providing a systematic framework for evaluating hybrid models on NISQ devices and identifying key areas for future research.
: Quantum physics-based quantum computing has gained attention as a new computing paradigm. As the quantum technology is advancing the Quantum Machine Learning (QML) is emerging as a new research area that combines quantum computing and machine learning. As classical machine learning continues to advance rapidly, by exploring new computing paradigms quantum machine learning is also progressing by taking a new parallel direction. In this paper, we implement and compare classical machine learning model, hybrid quantum-classical model, and fully quantum regression models for predicting Atmospheric Pollutant Concentration (APC). Multiple Linear Regression and Support Vector Regression are used as standard classical methods. Two hybrid models- Quantum Kernel Regression and Hybrid Quantum–Classical Regression, and a fully quantum model -Variational Quantum Regression are examined to understand how quantum techniques perform on today’s noisy quantum computers, known as Noisy Intermediate-Scale Quantum (NISQ) devices. The implementation and testing are conducted using a real time atmospheric pollutant concentration dataset containing data from different states of India, which is collected from Open Government Data (OGD) platform. The quantum models are implemented using quantum circuits and simulated using IBM Qiskit. The studies show that in this NISQ era- mainly limited qubits which are noisy and limited depth circuits, the classical regression models give more accurate predictions than quantum models. However, the quantum and hybrid approaches show comparable training times and provide useful insights into how these methods may scale in the future. Overall, this study highlights both the potential and the current limitations of quantum regression techniques when applied to real-world environmental prediction tasks in the NISQ era.