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

Encoding–Noise Interaction in Hybrid Quantum Classifiers: A Registered Simulation Study with Convergence Audit

Mohammed Ahmed

Published Oct 7, 2026• 0 citations
Publisher

Abstract

Comparisons of quantum data-encoding strategies are rarely accompanied by a registered estimand, an explicit treatment of where simulated noise is applied, or a classical reference point. This study reports a registered simulation study of the interaction between data encoding (angle versus ZZ feature map) and circuit noise (noiseless versus single-Pauli errors with probability p = 0.02 after every CNOT, plus readout flips) for an 8-qubit variational quantum classifier trained with SPSA. Three binary datasets, ten confirmatory seeds (240 runs) and circuit size as a secondary factor were used. Hyperparameters were tuned symmetrically on separate seeds and frozen before the confirmatory runs. The registered primary estimand, the Encoding x Noise interaction on validation balanced accuracy, was -0.154 (95% seed-cluster bootstrap CI [-0.181, -0.124]), which supports the registered directional hypothesis. ZZ-encoded models trained under noise met the pre-declared convergence criterion in 1 of 60 runs, against 48 of 60 for angle encoding, so training failure and signal degradation are not separated by the registered design. An exploratory follow-up, in which ZZ models were trained without noise and evaluated under noise, showed accuracy falling to 0.55-0.63 at p = 0.02, which points to the number of noisy two-qubit gates (56 additional CNOTs) as the main driver. The registered transfer hypothesis was not supported: robust selection did not differ from naive selection (+0.003, CI [-0.012, +0.016]). A class-balanced logistic regression exceeded every quantum configuration on validation accuracy. No quantum advantage is claimed.

DOI
10.5281/zenodo.23223354

Related papers

Rigorous Benchmarking of Variational Quantum Classifiers for Cancer Classification: Encoding Bandwidth and a Matched, Tuned Repeated Cross-Validation Study

B RAVI KIRAN PATNAIK

Variational quantum classifiers (VQCs) are often compared with classical models under unequal data partitions, tuning budgets, or statistical treatment. We benchmark a 4-qubit VQC against logistic regression, an RBF-SVM, random forest, a quantum-kernel SVM (QSVM), and a majority-class control on Breast Cancer Wisconsin (n=569), Coimbra (n=116), and Golub leukemia (n=72), using five stratified folds across five seeds (25 paired blocks per dataset), identical training-fold-fitted preprocessing, and Nadeau–Bengio corrected paired tests with Holm correction. A first arm with a frozen configuration and the common [0, π] angle encoding left the VQC at 0.563, 0.515, and 0.576 accuracy (majority 0.617, 0.552, 0.652). In a second arm, in which every model was tuned by the same inner cross-validation and the encoding scale was a tuned hyperparameter, VQC accuracy rose to 0.864, 0.616, and 0.747. On Wisconsin the VQC stayed significantly below every matched classical model (differences 0.079–0.097, Holm p≤0.006) but above the majority class; on Golub it was significantly below only the SVM (Holm p=0.016); on Coimbra no comparison was significant. A QSVM with the same feature map exceeded the VQC numerically on every dataset, and three optimizers reached the same training loss in pilot diagnostics, pointing to the encoding and the variational head rather than the optimizer. The Wisconsin arms also differ in sample size (120 versus 569). Results apply to this low-qubit, PCA-compressed, exactly simulated setting and do not indicate a quantum advantage.

•Oct 1, 2026•0 citations

Variational Quantum Classifiers: Comparing Data Encoding Strategies Under Noise

Shaurya Gera

Variational quantum classifiers (VQCs) are promising near-term quantum machine-learning models whose performance can depend on both the data-encoding method and circuit noise.This study compared angle encoding, amplitude encoding, and data re-uploading under three levels of simulated depolarizing noise using a fixed four-qubit, three-layer VQC.The comparison was repeated across the Iris versicolor-versus-virginica and Banknote Authentication datasets using multiple stratified data splits and parameter initializations, with a classical multilayer perceptron as a baseline.The relative performance of the encoding methods depended on the dataset: angle encoding produced the highest mean accuracy on Iris, while data re-uploading produced the highest mean accuracy on Banknote Authentication.Although statistical tests identified overall encoding differences in several Banknote conditions, no pairwise comparison remained significant after multiple-comparison correction.The classical baseline outperformed all tested quantum conditions.These findings show that encoding choice can affect VQC classifier behavior under simulated noise, but they do not establish a universally superior encoding method or direct robustness of the encoded quantum states.

Oxford Journal of Student Scholarship•Oct 1, 2026•0 citations

Where Does a Variational Quantum Classifier Lose Accuracy? Angle Encoding, Readout and the Fidelity-Kernel Reference in Cyberattack Detection

Angelos Thomos, Theodore Andronikos

When a variational quantum classifier performs poorly at detecting network attacks, it is hard to tell which part is to blame: the encoding, the trained circuit or the readout, which turns measurements into a class. In this work, we build upon a recent amplitude-encoded classifier, and we try to separate their contributions to the NSL-KDD benchmark in noiseless simulation. We compare angle and amplitude encoding, vary one component at a time, and measure each circuit against a classical yardstick: a support-vector machine on the encoded states, whose accuracy estimates what a well-trained circuit can reach. Most of the shortfall lies in the readout. Reading one qubit, or the parity of all qubits, leaves several percentage points unused, while a readout trained together with the circuit recovers most of them from the same measurements. This also explains why some classifiers stop improving as further layers are added. The encodings suit different needs. Angle encoding gives the more accurate detector on traffic like the training data and, from eight features up, shallower circuits; amplitude encoding needs far fewer qubits and catches more attacks of types never seen in training. Finally, we report that classical models, given the same inputs, are about as accurate as the best circuits, thus, registering no quantum advantage.

quant-pharXiv:2611.00885•Nov 3, 2026•0 citations

A Systematic Study of Noise Effects in Hybrid Quantum-Classical Machine Learning

Bhavna Bose, Muhammad Faryad

Near-term quantum machine learning (QML) models operate in environments wherein noise is unavoidable, arising from both imperfect classical data acquisition and the limitations of noisy intermediate-scale quantum (NISQ) hardware. Although most existing studies have focused primarily on quantum circuit noise in isolation, the combined influence of corrupted classical inputs and quantum hardware noise has received comparatively little attention. In this work, we present a systematic experimental study of the robustness of a variational quantum classifier under realistic multi-level noise conditions. Using the Titanic dataset as a benchmark, a range of dataset-level noise models-including speckle noise, impulse noise, quantization noise, and feature dropout are applied to classical features prior to quantum encoding using a ZZ feature map. In parallel, hardware-inspired quantum noise channels such as depolarizing noise, amplitude damping, phase damping, Pauli errors, and readout errors are incorporated at the circuit level using the Qiskit Aer simulator. The experimental results indicate that noise in classical input data can significantly intensify the effects of quantum decoherence, resulting in less stable training and noticeably lower classification accuracy. Together, these observations emphasize the importance of designing and evaluating quantum machine learning pipelines with noise in mind, and highlight the need to consider classical and quantum noise simultaneously when assessing QML performance in the NISQ era

quant-pharXiv:2604.11541•Apr 14, 2026•2 citations

Noise Resilience of Quantum Support Vector Machine with Selected Feature Maps

Muhammad Ahsan Shakeel, Saad Muzammil, Danyal Tayyub +1 more

Gate-level noise degrades the classification accuracy of Quantum Support Vector Machines (QSVMs) on Noisy Intermediate-Scale Quantum (NISQ) hardware, and the degree of degradation depends on how classical data is encoded into quantum states. We tested Z, ZZ, a Pauli, and an amplitude-inspired feature maps under depolarizing, bit-flip, and phase-flip noise channels in $52$ controlled experiments with error probabilities $p =0.01, 0.05, 0.10$, and $0.50$. The amplitude-inspired feature map had $100$\% test accuracy up to $p = 0.10$ across all three noise channels, while other feature maps fell to $65$-$90$\% under the same noise level and type. The Z feature map was found to be immune to phase-flip noise to a significantly high error rate, a consequence of the commutation relation $[R_Z, Z] = 0$. Entangled circuit variants produced generalization gaps in train-test sets of up to $17.5$\% under noise, whereas the amplitude variants maintained zero gap throughout. These results give practitioners data-driven criteria for a feature map on near-term quantum hardware.

quant-pharXiv:2608.17495•Aug 19, 2026•0 citations

Noise-Aware Evaluation of Hybrid Quantum–Classical Models for Medical Binary Classification: Training Time vs Noise Level Analysis

Abhishek Gajanan Kolekar

This study presents a noise-aware evaluation framework for hybrid quantum–classical machine learning models applied to medical binary classification. Using three healthcare datasets—Pima Indians Diabetes, Heart Disease, and Breast Cancer Wisconsin—the performance of a Variational Quantum Classifier (VQC) was analyzed under multiple depolarizing noise levels (0.00–0.10). Classical baselines, Logistic Regression (LR) and Support Vector Machine (SVM), were implemented for comparison. The research investigates the impact of noise on predictive accuracy, F1-score, and training time, highlighting the computational overhead of hybrid quantum–classical models. Zero Noise Extrapolation (ZNE), using linear and Richardson methods, was applied to mitigate noise effects, revealing that while ZNE improves performance under certain conditions, extrapolation instability can occur. Transpiler optimization also contributed to improved VQC stability by reducing circuit complexity. Results show that classical models consistently outperformed the VQC in accuracy, stability, and training efficiency. VQC exhibited fluctuating performance under noise, requiring significantly higher training time, and increasing SPSA iterations did not always yield stable accuracy improvements. These findings underscore the challenges of noise robustness, optimization instability, and computational cost in the NISQ era, and suggest directions for future work, including alternative optimizers, adaptive ansatz design, advanced error mitigation, and real hardware benchmarking. This research provides a systematic benchmarking framework for evaluating training time versus noise-level trade-offs, contributing to the development of more reliable quantum machine learning models for medical applications. Keywords: Quantum Machine Learning, Variational Quantum Classifier, Depolarizing Noise, Zero Noise Extrapolation, Medical Classification, Hybrid Quantum–Classical Models, Training Time Analysis

•Apr 3, 2026•0 citations
Comparisons of quantum data-encoding strategies are rarely accompanied by a registered estimand, an explicit treatment of where simulated noise is applied, or a classical reference point. This study reports a registered simulation study of the interaction between data encoding (angle versus ZZ feature map) and circuit noise (noiseless versus single-Pauli errors with probability p = 0.02 after every CNOT, plus readout flips) for an 8-qubit variational quantum classifier trained with SPSA. Three binary datasets, ten confirmatory seeds (240 runs) and circuit size as a secondary factor were used. Hyperparameters were tuned symmetrically on separate seeds and frozen before the confirmatory runs. The registered primary estimand, the Encoding x Noise interaction on validation balanced accuracy, was -0.154 (95% seed-cluster bootstrap CI [-0.181, -0.124]), which supports the registered directional hypothesis. ZZ-encoded models trained under noise met the pre-declared convergence criterion in 1 of 60 runs, against 48 of 60 for angle encoding, so training failure and signal degradation are not separated by the registered design. An exploratory follow-up, in which ZZ models were trained without noise and evaluated under noise, showed accuracy falling to 0.55-0.63 at p = 0.02, which points to the number of noisy two-qubit gates (56 additional CNOTs) as the main driver. The registered transfer hypothesis was not supported: robust selection did not differ from naive selection (+0.003, CI [-0.012, +0.016]). A class-balanced logistic regression exceeded every quantum configuration on validation accuracy. No quantum advantage is claimed.
Variational quantum classifiers (VQCs) are often compared with classical models under unequal data partitions, tuning budgets, or statistical treatment. We benchmark a 4-qubit VQC against logistic regression, an RBF-SVM, random forest, a quantum-kernel SVM (QSVM), and a majority-class control on Breast Cancer Wisconsin (n=569), Coimbra (n=116), and Golub leukemia (n=72), using five stratified folds across five seeds (25 paired blocks per dataset), identical training-fold-fitted preprocessing, and Nadeau–Bengio corrected paired tests with Holm correction. A first arm with a frozen configuration and the common [0, π] angle encoding left the VQC at 0.563, 0.515, and 0.576 accuracy (majority 0.617, 0.552, 0.652). In a second arm, in which every model was tuned by the same inner cross-validation and the encoding scale was a tuned hyperparameter, VQC accuracy rose to 0.864, 0.616, and 0.747. On Wisconsin the VQC stayed significantly below every matched classical model (differences 0.079–0.097, Holm p≤0.006) but above the majority class; on Golub it was significantly below only the SVM (Holm p=0.016); on Coimbra no comparison was significant. A QSVM with the same feature map exceeded the VQC numerically on every dataset, and three optimizers reached the same training loss in pilot diagnostics, pointing to the encoding and the variational head rather than the optimizer. The Wisconsin arms also differ in sample size (120 versus 569). Results apply to this low-qubit, PCA-compressed, exactly simulated setting and do not indicate a quantum advantage.
Variational quantum classifiers (VQCs) are promising near-term quantum machine-learning models whose performance can depend on both the data-encoding method and circuit noise.This study compared angle encoding, amplitude encoding, and data re-uploading under three levels of simulated depolarizing noise using a fixed four-qubit, three-layer VQC.The comparison was repeated across the Iris versicolor-versus-virginica and Banknote Authentication datasets using multiple stratified data splits and parameter initializations, with a classical multilayer perceptron as a baseline.The relative performance of the encoding methods depended on the dataset: angle encoding produced the highest mean accuracy on Iris, while data re-uploading produced the highest mean accuracy on Banknote Authentication.Although statistical tests identified overall encoding differences in several Banknote conditions, no pairwise comparison remained significant after multiple-comparison correction.The classical baseline outperformed all tested quantum conditions.These findings show that encoding choice can affect VQC classifier behavior under simulated noise, but they do not establish a universally superior encoding method or direct robustness of the encoded quantum states.
When a variational quantum classifier performs poorly at detecting network attacks, it is hard to tell which part is to blame: the encoding, the trained circuit or the readout, which turns measurements into a class. In this work, we build upon a recent amplitude-encoded classifier, and we try to separate their contributions to the NSL-KDD benchmark in noiseless simulation. We compare angle and amplitude encoding, vary one component at a time, and measure each circuit against a classical yardstick: a support-vector machine on the encoded states, whose accuracy estimates what a well-trained circuit can reach. Most of the shortfall lies in the readout. Reading one qubit, or the parity of all qubits, leaves several percentage points unused, while a readout trained together with the circuit recovers most of them from the same measurements. This also explains why some classifiers stop improving as further layers are added. The encodings suit different needs. Angle encoding gives the more accurate detector on traffic like the training data and, from eight features up, shallower circuits; amplitude encoding needs far fewer qubits and catches more attacks of types never seen in training. Finally, we report that classical models, given the same inputs, are about as accurate as the best circuits, thus, registering no quantum advantage.
Near-term quantum machine learning (QML) models operate in environments wherein noise is unavoidable, arising from both imperfect classical data acquisition and the limitations of noisy intermediate-scale quantum (NISQ) hardware. Although most existing studies have focused primarily on quantum circuit noise in isolation, the combined influence of corrupted classical inputs and quantum hardware noise has received comparatively little attention. In this work, we present a systematic experimental study of the robustness of a variational quantum classifier under realistic multi-level noise conditions. Using the Titanic dataset as a benchmark, a range of dataset-level noise models-including speckle noise, impulse noise, quantization noise, and feature dropout are applied to classical features prior to quantum encoding using a ZZ feature map. In parallel, hardware-inspired quantum noise channels such as depolarizing noise, amplitude damping, phase damping, Pauli errors, and readout errors are incorporated at the circuit level using the Qiskit Aer simulator. The experimental results indicate that noise in classical input data can significantly intensify the effects of quantum decoherence, resulting in less stable training and noticeably lower classification accuracy. Together, these observations emphasize the importance of designing and evaluating quantum machine learning pipelines with noise in mind, and highlight the need to consider classical and quantum noise simultaneously when assessing QML performance in the NISQ era
Gate-level noise degrades the classification accuracy of Quantum Support Vector Machines (QSVMs) on Noisy Intermediate-Scale Quantum (NISQ) hardware, and the degree of degradation depends on how classical data is encoded into quantum states. We tested Z, ZZ, a Pauli, and an amplitude-inspired feature maps under depolarizing, bit-flip, and phase-flip noise channels in 525252 controlled experiments with error probabilities p=0.01,0.05,0.10p =0.01, 0.05, 0.10p=0.01,0.05,0.10, and 0.500.500.50. The amplitude-inspired feature map had 100100100\% test accuracy up to p=0.10p = 0.10p=0.10 across all three noise channels, while other feature maps fell to 656565-909090\% under the same noise level and type. The Z feature map was found to be immune to phase-flip noise to a significantly high error rate, a consequence of the commutation relation [RZ,Z]=0[R_Z, Z] = 0[RZ​,Z]=0. Entangled circuit variants produced generalization gaps in train-test sets of up to 17.517.517.5\% under noise, whereas the amplitude variants maintained zero gap throughout. These results give practitioners data-driven criteria for a feature map on near-term quantum hardware.
This study presents a noise-aware evaluation framework for hybrid quantum–classical machine learning models applied to medical binary classification. Using three healthcare datasets—Pima Indians Diabetes, Heart Disease, and Breast Cancer Wisconsin—the performance of a Variational Quantum Classifier (VQC) was analyzed under multiple depolarizing noise levels (0.00–0.10). Classical baselines, Logistic Regression (LR) and Support Vector Machine (SVM), were implemented for comparison. The research investigates the impact of noise on predictive accuracy, F1-score, and training time, highlighting the computational overhead of hybrid quantum–classical models. Zero Noise Extrapolation (ZNE), using linear and Richardson methods, was applied to mitigate noise effects, revealing that while ZNE improves performance under certain conditions, extrapolation instability can occur. Transpiler optimization also contributed to improved VQC stability by reducing circuit complexity. Results show that classical models consistently outperformed the VQC in accuracy, stability, and training efficiency. VQC exhibited fluctuating performance under noise, requiring significantly higher training time, and increasing SPSA iterations did not always yield stable accuracy improvements. These findings underscore the challenges of noise robustness, optimization instability, and computational cost in the NISQ era, and suggest directions for future work, including alternative optimizers, adaptive ansatz design, advanced error mitigation, and real hardware benchmarking. This research provides a systematic benchmarking framework for evaluating training time versus noise-level trade-offs, contributing to the development of more reliable quantum machine learning models for medical applications. Keywords: Quantum Machine Learning, Variational Quantum Classifier, Depolarizing Noise, Zero Noise Extrapolation, Medical Classification, Hybrid Quantum–Classical Models, Training Time Analysis