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

Trainability-Aware Quantum Gate Pruning for Variational Quantum Circuits Using Saliency and Quantum Fisher Information

M Vedhanth, S Mahadevi, Anil G Kumar

Published Jan 1, 2026• 0 citations
Publisher

Abstract

Variational quantum circuits (VQCs) suffer from barren plateaus, where gradients vanish exponentially with system size, limiting effective optimization. While prior mitigation strategies focus on circuit design and initialization, the effect of circuit compression – specifically, which gates are removed – on trainability remains comparatively unexplored. In this work we study quantum gate pruning from a trainability perspective in a controlled, noiseless 6-qubit simulation testbed, and report numerical evidence that gate-selection criterion plays a substantial role in preserving the optimization landscape of compressed circuits. We propose Q-AGASP, a saliency-based pruning framework that ranks parameterized quantum gates using a combination of gradient-weight sensitivity and quantum activation, and introduce a Quantum Fisher Information (QFI)-motivated saliency variant that targets parameters associated with low gradient variance. Both approaches are applied through an iterative pruning schedule with frozen weights. Experiments on a hybrid quantum-classical model (6-qubit, 144-parameter VQC) show that saliency-guided pruning improves the gradient variance of the surviving parameters relative to uninformed methods – by 6–9 $\times $ at 90% gate sparsity in our tested regime – while random or magnitude-based pruning can produce near-zero gradient variance in some seeds. We further find that post-fine-tuning accuracy converges across methods under our 15-epoch fine-tuning budget, which suggests that accuracy and pre-fine-tuning gradient variance are complementary rather than interchangeable diagnostics, with accuracy becoming less discriminative precisely when the fine-tuning budget is generous relative to circuit size. To analyze how gate rankings evolve during compression we introduce saliency drift, a Spearman-rank-correlation-based metric, and derive a distribution-free bound relating the observed correlation to the magnitude of ranking reshuffles; drift increases with sparsity in our experiments, consistent with iterative recomputation capturing structural changes not visible to one-shot pruning. We are explicit that these findings are numerical observations from a single ansatz, qubit count, and dataset rather than architecture-independent theorems, and we discuss how ansatz topology, entanglement structure, and circuit depth would be expected to modulate them, together with the changes needed to extend the framework to noisy hardware and to chemistry-motivated ansätze.

DOI
10.1109/access.2026.3742327

Related papers

Variational Quantum Machine Learning for Binary Classification on NISQ Hardware: Expressibility, Trainability, and Noise Robustness

Priya S. Mehta, Aleksandr V. Korolev

This study investigates classification performance and trainability of variational quantum classifiers on near-term noisy quantum hardware across circuit depth and qubit count regimes within the context of quantum computing and machine learning, an area of growing scientific importance given its implications for hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment. Using gradient-based parameter optimization of parameterized quantum circuits with hardware noise modeling via density matrix simulation and physical execution on IBM Eagle processors, we examine expressibility of parameterized circuits enabling nonlinear decision boundaries in feature Hilbert space, with trainability constrained by barren plateau gradients at large qubit counts in 96 circuit architectures (4 qubit counts x 4 depths x 6 ansatz families) evaluated across 5 classification datasets with 20 random initialization seeds each drawn from IBM Quantum Eagle (127-qubit) and Falcon (27-qubit) processors with noise mitigation via zero-noise extrapolation. Results indicate that 10-qubit, 4-layer VQC achieves 84.2% classification accuracy on the best benchmark, within 3.8 percentage points of classical SVM, but gradient variance decays exponentially with qubit count (variance halving per 2 added qubits), confirming barren plateau scaling (p < 0.001), with 84.2% accuracy, 3.8 pp gap to classical SVM baseline as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to quantum computing and machine learning and carry actionable implications for the design of programs and policies targeting hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment.

Princeton Journal of Pre-Collegiate Research•Jan 1, 2021•0 citations

Emergent Effective Depth in Quantum Machine Learning: Empirical Gradient Attenuation Scaling in Noisy Variational Quantum Algorithms

Carlos Mario Braga, Manuel A. Serrano, Eduardo Fernández‐Medina

Noisy Intermediate-Scale Quantum (NISQ) devices impose structural constraints on the training of Variational Quantum Algorithms (VQAs). In the absence of full error correction, each quantum gate introduces a non-zero probability of error that accumulates with circuit depth, while gradient estimation through finite sampling adds additional statistical variability. As a result, convergence depends on a delicate balance between physical coherence and estimation variance. In this setting, the depth of the variational ansatz not only determines model expressivity, but also its operational feasibility under noise. Without an explicit characterization of the interaction between error accumulation, number of measurements (shots), and control strategies, increasing experimental resources may fail to improve performance and can even become counterproductive. In this work, we empirically analyze how circuit depth affects the practical signal of the gradient under depolarizing noise. Across multiple configurations ($q=4-7$qubits), we observe a pattern consistent with an effective exponential attenuation of the coherent gradient contribution, characterized by a per-layer rate$\epsilon_{\text{eff}}$. The estimated magnitude of$\epsilon_{\text{eff}}$depends on circuit complexity and remains largely independent of the measurement budget, indicating that it primarily reflects architectural and physical factors rather than sampling effects. The analysis of adaptive control strategies further suggests that accuracy gains are constrained by gradient attenuation, rather than scaling significantly with circuit depth. However, a crossmode comparison across three distinct training strategies reveals that, while accuracy improvements remain modest, adaptive perturbations substantially improve the statistical detectability of the attenuation signal, acting as diagnostic probes rather than mere optimization heuristics. Taken together, these results suggest that the usable depth of a variational ansatz can be interpreted as an emergent property of the effective signal-to-noise regime, governed by a measurable structural parameter whose ranking across circuit configurations is preserved independently of the training mode. This framework enables principled comparison of circuit architectures in terms of their effective trainability under noise, providing guidance for architectural and control design in Quantum Machine Learning under NISQ conditions.

IEEE International Conference on Quantum Software (QSW)•Jul 1, 2026•0 citations

Predictingthe Trainability of Variational Quantum Circuits: A Data-Driven Model for BarrenPlateaus

Md Habibur Rahman

Abstract Barren plateaus are the main obstacle to training quantum machine-learning models. On a barren plateau, the cost-function gradient shrinks exponentially as the circuit grows, and training stalls. Today the only reliable test is to measure it. That means sampling many gradients for every candidate circuit, which is expensive. We ask a different question. Can a classical model predict trainability from the circuit architecture alone? The architecture here means qubit count, depth, ansatz type, entanglement pattern, entangler gate, and cost-observable locality. We test this directly. We use an exact statevector simulator to build a dataset of \Nrows{} random variational circuits. For each circuit we estimate the gradient variance $\mathrm{Var}[\partial C/\partial\theta]$ from \Samples{} random parameter vectors, using the parameter-shift rule. This variance is the label. A gradient-boosted regressor then predicts $\log_{10}\mathrm{Var}$ with $R^2 = \RsqHold{}$ on held-out architectures. A classifier separates barren from trainable circuits with an area under the ROC curve of \AucHold{}. We also test transfer to larger circuits. We train only on small circuits and test on the two largest unseen sizes ($n=11,12$). The model still reaches $R^2 = \RsqExtrap{}$. Permutation importance then recovers established theory. Cost-observable locality and entanglement structure matter most. We make no quantum-advantage claim. This is classical machine learning about quantum circuits. It offers a cheap way to screen ans\"atze, and it shows which design choices cause untrainability.

•Aug 19, 2026•0 citations

Trainability and Expressivity of Hamming-Weight Preserving Quantum Circuits for Machine Learning

Léo Monbroussou, Eliott Z. Mamon, Jonas Landman +3 more

Quantum machine learning (QML) has become a promising area for real world applications of quantum computers, but near-term methods and their scalability are still important research topics. In this context, we analyze the trainability and controllability of specific Hamming weight preserving variational quantum circuits (VQCs). These circuits use qubit gates that preserve subspaces of the Hilbert space, spanned by basis states with fixed Hamming weight k . In this work, we first design and prove the feasibility of new heuristic data loaders, performing quantum amplitude encoding of (nk) -dimensional vectors by training an n -qubit quantum circuit. These data loaders are obtained using controllability arguments, by checking the Quantum Fisher Information Matrix (QFIM)'s rank. Second, we provide a theoretical justification for the fact that the rank of the QFIM of any VQC state is almost-everywhere constant, which is of separate interest. Lastly, we analyze the trainability of Hamming weight preserving circuits, and show that the variance of the l2 cost function gradient is bounded according to the dimension (nk) of the subspace. This proves conditions of existence/lack of Barren Plateaus for these circuits, and highlights a setting where a recent conjecture on the link between controllability and trainability of variational quantum circuits does not apply.

arXiv:2309.15547•Quantum•May 15, 2025•43 citations

Late Breaking Results: Hardware-Aware Compilation Reshapes Trainability in Variational Quantum Circuits

Muhammad Kashif, Muhammad Shafique

Variational quantum circuits (VQCs) are typically evaluated at the logical design level when analyzing trainability. However, execution on real quantum devices requires hardware-aware compilation (transpilation) to satisfy qubit connectivity and native gate-set constraints. In this paper, we examine how transpilation can alter the gradient statistics. Using parameter-shift differentiation and gradient variance estimation, we compare logical and transpiled circuits across three representative ansatz families: EfficientSU2 (dense entanglement), TTN (tree tensor network), and RealAmplitudes (linear entanglement). We observe architecture-dependent trainability shifts where densely entangling circuits exhibit pronounced gradient reshaping in shallow regimes, structured tensor-network circuits remain comparatively robust, and linear architectures show mixed behavior. Deep circuits across all families display minimal sensitivity to hardware-aware compilation. These findings demonstrate that transpilation acts as an implicit structural transformation of the optimization landscape, motivating compilation-aware analysis and co-design for VQCs.

quant-pharXiv:2604.16527•ACM/IEEE Design Automation Conference (DAC)•Apr 17, 2026•2 citations

Parameter Efficiency of Variational Quantum Circuits in Image Classification: The Role of Circuit Depth

Pranauv Muthuraman

The practical value of quantum machine learning must be derived from efficiency rather than scale, as current NISQ processors support only small, shallow circuits. This paper evaluated whether variational quantum circuits (VQCs) can reduce the trainable parameter count in an image classification scenario while preserving accuracy. Experimentation compared quantum classifier heads to parameter-matched classical heads on an identical set of frozen features drawn from the Fashion-MNIST dataset by a CNN. 9 configurations of a quantum circuit crossed 4, 6, and 8 qubits with 1, 2, and 4 entangling layers, each matched to a multilayer perceptron with an approximately equal parameter count. The linear probes served as a floor baseline for all 105 runs across 5 random seed values. 7 out of 9 quantum-based configurations met the mean test accuracy of no more than 1 percentage point below its matched classical head, enforced by the pre-specified criterion. The 2 failures were single-layer circuits, with each falling roughly 8 points below its matched classical head and 12 points below the circuit-free linear probe; in total, 6 configs with 2+ layers passed, with the smallest at 74 trainable parameters. The general trend noticed was that accuracy increased progressively with depth. The 4-layer circuits exceeded the linear probe at every qubit count and produced the highest accuracy at 6 and 8 qubits. The simulation ran noiselessly; VQCs preserved accuracy at the matched parameter budget when depth was given, showcasing depth rather than qubit count gating performance.

•Aug 15, 2026•0 citations
Variational quantum circuits (VQCs) suffer from barren plateaus, where gradients vanish exponentially with system size, limiting effective optimization. While prior mitigation strategies focus on circuit design and initialization, the effect of circuit compression – specifically, which gates are removed – on trainability remains comparatively unexplored. In this work we study quantum gate pruning from a trainability perspective in a controlled, noiseless 6-qubit simulation testbed, and report numerical evidence that gate-selection criterion plays a substantial role in preserving the optimization landscape of compressed circuits. We propose Q-AGASP, a saliency-based pruning framework that ranks parameterized quantum gates using a combination of gradient-weight sensitivity and quantum activation, and introduce a Quantum Fisher Information (QFI)-motivated saliency variant that targets parameters associated with low gradient variance. Both approaches are applied through an iterative pruning schedule with frozen weights. Experiments on a hybrid quantum-classical model (6-qubit, 144-parameter VQC) show that saliency-guided pruning improves the gradient variance of the surviving parameters relative to uninformed methods – by 6–9 ×\times × at 90% gate sparsity in our tested regime – while random or magnitude-based pruning can produce near-zero gradient variance in some seeds. We further find that post-fine-tuning accuracy converges across methods under our 15-epoch fine-tuning budget, which suggests that accuracy and pre-fine-tuning gradient variance are complementary rather than interchangeable diagnostics, with accuracy becoming less discriminative precisely when the fine-tuning budget is generous relative to circuit size. To analyze how gate rankings evolve during compression we introduce saliency drift, a Spearman-rank-correlation-based metric, and derive a distribution-free bound relating the observed correlation to the magnitude of ranking reshuffles; drift increases with sparsity in our experiments, consistent with iterative recomputation capturing structural changes not visible to one-shot pruning. We are explicit that these findings are numerical observations from a single ansatz, qubit count, and dataset rather than architecture-independent theorems, and we discuss how ansatz topology, entanglement structure, and circuit depth would be expected to modulate them, together with the changes needed to extend the framework to noisy hardware and to chemistry-motivated ansätze.
This study investigates classification performance and trainability of variational quantum classifiers on near-term noisy quantum hardware across circuit depth and qubit count regimes within the context of quantum computing and machine learning, an area of growing scientific importance given its implications for hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment. Using gradient-based parameter optimization of parameterized quantum circuits with hardware noise modeling via density matrix simulation and physical execution on IBM Eagle processors, we examine expressibility of parameterized circuits enabling nonlinear decision boundaries in feature Hilbert space, with trainability constrained by barren plateau gradients at large qubit counts in 96 circuit architectures (4 qubit counts x 4 depths x 6 ansatz families) evaluated across 5 classification datasets with 20 random initialization seeds each drawn from IBM Quantum Eagle (127-qubit) and Falcon (27-qubit) processors with noise mitigation via zero-noise extrapolation. Results indicate that 10-qubit, 4-layer VQC achieves 84.2% classification accuracy on the best benchmark, within 3.8 percentage points of classical SVM, but gradient variance decays exponentially with qubit count (variance halving per 2 added qubits), confirming barren plateau scaling (p < 0.001), with 84.2% accuracy, 3.8 pp gap to classical SVM baseline as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to quantum computing and machine learning and carry actionable implications for the design of programs and policies targeting hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment.
Noisy Intermediate-Scale Quantum (NISQ) devices impose structural constraints on the training of Variational Quantum Algorithms (VQAs). In the absence of full error correction, each quantum gate introduces a non-zero probability of error that accumulates with circuit depth, while gradient estimation through finite sampling adds additional statistical variability. As a result, convergence depends on a delicate balance between physical coherence and estimation variance. In this setting, the depth of the variational ansatz not only determines model expressivity, but also its operational feasibility under noise. Without an explicit characterization of the interaction between error accumulation, number of measurements (shots), and control strategies, increasing experimental resources may fail to improve performance and can even become counterproductive. In this work, we empirically analyze how circuit depth affects the practical signal of the gradient under depolarizing noise. Across multiple configurations (q=4−7q=4-7q=4−7qubits), we observe a pattern consistent with an effective exponential attenuation of the coherent gradient contribution, characterized by a per-layer rateϵeff\epsilon_{\text{eff}}ϵeff​. The estimated magnitude ofϵeff\epsilon_{\text{eff}}ϵeff​depends on circuit complexity and remains largely independent of the measurement budget, indicating that it primarily reflects architectural and physical factors rather than sampling effects. The analysis of adaptive control strategies further suggests that accuracy gains are constrained by gradient attenuation, rather than scaling significantly with circuit depth. However, a crossmode comparison across three distinct training strategies reveals that, while accuracy improvements remain modest, adaptive perturbations substantially improve the statistical detectability of the attenuation signal, acting as diagnostic probes rather than mere optimization heuristics. Taken together, these results suggest that the usable depth of a variational ansatz can be interpreted as an emergent property of the effective signal-to-noise regime, governed by a measurable structural parameter whose ranking across circuit configurations is preserved independently of the training mode. This framework enables principled comparison of circuit architectures in terms of their effective trainability under noise, providing guidance for architectural and control design in Quantum Machine Learning under NISQ conditions.
Abstract Barren plateaus are the main obstacle to training quantum machine-learning models. On a barren plateau, the cost-function gradient shrinks exponentially as the circuit grows, and training stalls. Today the only reliable test is to measure it. That means sampling many gradients for every candidate circuit, which is expensive. We ask a different question. Can a classical model predict trainability from the circuit architecture alone? The architecture here means qubit count, depth, ansatz type, entanglement pattern, entangler gate, and cost-observable locality. We test this directly. We use an exact statevector simulator to build a dataset of \Nrows{} random variational circuits. For each circuit we estimate the gradient variance Var[∂C/∂θ]\mathrm{Var}[\partial C/\partial\theta]Var[∂C/∂θ] from \Samples{} random parameter vectors, using the parameter-shift rule. This variance is the label. A gradient-boosted regressor then predicts log⁡10Var\log_{10}\mathrm{Var}log10​Var with $R^2 = \RsqHold{}$ on held-out architectures. A classifier separates barren from trainable circuits with an area under the ROC curve of \AucHold{}. We also test transfer to larger circuits. We train only on small circuits and test on the two largest unseen sizes (n=11,12n=11,12n=11,12). The model still reaches $R^2 = \RsqExtrap{}$. Permutation importance then recovers established theory. Cost-observable locality and entanglement structure matter most. We make no quantum-advantage claim. This is classical machine learning about quantum circuits. It offers a cheap way to screen ans\"atze, and it shows which design choices cause untrainability.
Quantum machine learning (QML) has become a promising area for real world applications of quantum computers, but near-term methods and their scalability are still important research topics. In this context, we analyze the trainability and controllability of specific Hamming weight preserving variational quantum circuits (VQCs). These circuits use qubit gates that preserve subspaces of the Hilbert space, spanned by basis states with fixed Hamming weight k . In this work, we first design and prove the feasibility of new heuristic data loaders, performing quantum amplitude encoding of (nk) -dimensional vectors by training an n -qubit quantum circuit. These data loaders are obtained using controllability arguments, by checking the Quantum Fisher Information Matrix (QFIM)'s rank. Second, we provide a theoretical justification for the fact that the rank of the QFIM of any VQC state is almost-everywhere constant, which is of separate interest. Lastly, we analyze the trainability of Hamming weight preserving circuits, and show that the variance of the l2 cost function gradient is bounded according to the dimension (nk) of the subspace. This proves conditions of existence/lack of Barren Plateaus for these circuits, and highlights a setting where a recent conjecture on the link between controllability and trainability of variational quantum circuits does not apply.
Variational quantum circuits (VQCs) are typically evaluated at the logical design level when analyzing trainability. However, execution on real quantum devices requires hardware-aware compilation (transpilation) to satisfy qubit connectivity and native gate-set constraints. In this paper, we examine how transpilation can alter the gradient statistics. Using parameter-shift differentiation and gradient variance estimation, we compare logical and transpiled circuits across three representative ansatz families: EfficientSU2 (dense entanglement), TTN (tree tensor network), and RealAmplitudes (linear entanglement). We observe architecture-dependent trainability shifts where densely entangling circuits exhibit pronounced gradient reshaping in shallow regimes, structured tensor-network circuits remain comparatively robust, and linear architectures show mixed behavior. Deep circuits across all families display minimal sensitivity to hardware-aware compilation. These findings demonstrate that transpilation acts as an implicit structural transformation of the optimization landscape, motivating compilation-aware analysis and co-design for VQCs.
The practical value of quantum machine learning must be derived from efficiency rather than scale, as current NISQ processors support only small, shallow circuits. This paper evaluated whether variational quantum circuits (VQCs) can reduce the trainable parameter count in an image classification scenario while preserving accuracy. Experimentation compared quantum classifier heads to parameter-matched classical heads on an identical set of frozen features drawn from the Fashion-MNIST dataset by a CNN. 9 configurations of a quantum circuit crossed 4, 6, and 8 qubits with 1, 2, and 4 entangling layers, each matched to a multilayer perceptron with an approximately equal parameter count. The linear probes served as a floor baseline for all 105 runs across 5 random seed values. 7 out of 9 quantum-based configurations met the mean test accuracy of no more than 1 percentage point below its matched classical head, enforced by the pre-specified criterion. The 2 failures were single-layer circuits, with each falling roughly 8 points below its matched classical head and 12 points below the circuit-free linear probe; in total, 6 configs with 2+ layers passed, with the smallest at 74 trainable parameters. The general trend noticed was that accuracy increased progressively with depth. The 4-layer circuits exceeded the linear probe at every qubit count and produced the highest accuracy at 6 and 8 qubits. The simulation ran noiselessly; VQCs preserved accuracy at the matched parameter budget when depth was given, showcasing depth rather than qubit count gating performance.