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

arXiv:2508.05559 · Physical Review Research

Design of expressive and trainable pulse-based quantum machine learning models

Han-Xiao Tao, Xin Wang, Re-Bing Wu

Published Feb 9, 2026• 4 citations
arXivHTMLPDF

Abstract

Pulse-based quantum machine learning (QML) has emerged as a promising paradigm in quantum artificial intelligence due to its exceptional hardware efficiency. For real-world applications, it is crucial that pulse-based models are both expressive and trainable. Previous studies have shown that pulse-based models operating under dynamical symmetry can be effectively trained, benefiting from a favorable loss landscape devoid of barren plateaus. However, this approach may lead to uncontrollability, which can limit expressivity if the model is poorly designed. In this paper, we explore the conditions necessary for pulse-based QML models to maintain both expressivity and trainability. We present a key requirement involving the system's initial state, the measurement observable, and the underlying dynamical symmetry Lie algebra, supported by numerical simulations. Our results provide a comprehensive framework for the design of practical pulse-based QML models that strike a good balance between expressivity and trainability.

DOI
10.1103/gvj4-kqwd

Related papers

On the Role of Controllability in Pulse-based Quantum Machine Learning Models

Han-Xiao Tao, Re-Bing Wu

Pulse-based quantum machine learning (QML) models possess full expressivity when they are ensemble controllable. However, the accompanied barren plateaus render training intractable for high-dimensional systems. In this paper, we show that the trade-off can be understood via controllability analysis. We first apply the Fliess-series expansion to pulse-based QML models to investigate the effect of control system structure on model expressivity, which leads to a universal criterion for assessing the expressivity of generic QML models. We then show that uncontrollable pulse-based models that evolve on low-dimensional manifolds can achieve better balance between expressivity and trainability. Numerical experiments verify the proposed criterion and our analysis, which further demonstrate that the expressivity can be enhanced by increasing dimensionality, while barren plateaus can be avoided by reducing the controllability. Our approach provides a promising path for designing pulse-based QML models that are both highly expressive and trainable.

arXiv:2405.09135•IEEE Conference on Decision and Control (CDC)•Dec 16, 2024•2 citations

Quantum Machine Learning for Industrial Applications

Léo Monbroussou

Recent advances in Machine Learning have transformed numerous industrial sectors, yet classical paradigms face fundamental limitations: rapidly growing data volumes, rising computational costs, significant energy consumption, and the physical scaling limits of conventional hardware architectures. Quantum computing has emerged as a promising computational paradigm to address these challenges, giving rise to the field of Quantum Machine Learning (QML). In this thesis, the theoretical foundations of QML are investigated, with a focus on near-term and future practical applications. Three central challenges are addressed: the trainability of variational quantum circuits, their expressivity, and their resistance to efficient classical simulation. The trainability of Hamming-weight preserving variational quantum circuits is first studied, and theoretical guarantees are established that resolve an open conjecture on the absence of barren plateaus for this circuit family. Subspace-preserving QML algorithms are then introduced, including photonic circuits and quantum convolutional neural networks, and are designed to mimic classical ML subroutines while offering polynomial quantum advantage. Finally, variational quantum circuits are analyzed as quantum Fourier models, and a framework is derived to jointly characterize expressivity and trainability, from which conditions are obtained under which quantum models provably separate from their classical counterparts. These contributions are intended to advance the theoretical roadmap for harnessing near-term and future quantum technologies in real-world applications.

quant-pharXiv:2606.14822•Jun 15, 2026•1 citation

Provably Trainable Rotationally Equivariant Quantum Machine Learning

Maxwell T. West, Jamie Heredge, Martin E. Sevior +1 more

Exploiting the power of quantum computation to realize superior machine learning algorithms has been a major research focus of recent years, but the prospects of quantum machine learning (QML) remain dampened by considerable technical challenges. A particularly significant issue is that generic QML models suffer from so-called barren plateaus in their training landscapes—large regions where cost function gradients vanish exponentially in the number of qubits employed, rendering large models effectively untrainable. A leading strategy for combating this effect is to build problem-specific models that take into account the symmetries of their data in order to focus on a smaller, relevant subset of Hilbert space. In this work, we introduce a family of rotationally equivariant QML models built upon the quantum Fourier transform, and leverage recent insights from the Lie-algebraic study of QML models to prove that (a subset of) our models do not exhibit barren plateaus. In addition to our analytical results we numerically test our rotationally equivariant models on a dataset of simulated scanning tunneling microscope images of phosphorus impurities in silicon, where rotational symmetry naturally arises, and find that they dramatically outperform their generic counterparts in practice. Published by the American Physical Society 2024

arXiv:2311.05873•PRX Quantum•Jul 31, 2024•54 citations

Beyond the Expressivity-Trainability Paradox: A Dynamical Lie Algebra Perspective on Navigating Barren Plateaus in Quantum Machine Learning

Kung-Ming Lan, Edward Huang

As Quantum Machine Learning (QML) transitions toward practical implementation, the field faces a critical architectural bottleneck that challenges the fundamental assumptions of classical statistical learning theory. In classical deep learning, increasing model capacity typically risks overfitting. However, this study advances a counter-intuitive paradigm: unstructured contemporary QML architectures suffer from a profound state of quantum underfitting, driven by the "expressivity-trainability paradox." We demonstrate that the vast Hilbert space capacity of Parameterized Quantum Circuits (PQCs)-traditionally chased as the source of quantum advantage is the direct mathematical cause of Barren Plateaus (BPs), where gradient landscapes become exponentially flat. By synthesizing recent breakthroughs in Dynamical Lie Algebras (DLAs) and Geometric QML, we establish a comprehensive framework linking the algebraic dimension of circuit generators to their optimization dynamics. Furthermore, we empirically validate this framework on a non-linear binary classification task, illuminating a uniquely quantum manifestation of the bias-variance tradeoff: while unstructured architectures achieve near-perfect training accuracy via unscalable parameterization (quantum overfitting), embedding group-theoretic geometric priors acts as a structural regularizer. By restricting the DLA growth to a polynomial regime, our symmetry-preserving approach sacrifices raw memorization capacity to guarantee scalable, gradient-rich training landscapes, offering a robust roadmap for "Trainability-by-Design" in scalable quantum neural networks.

cs.LGCross-listedarXiv:2606.31536•Jul 1, 2026•1 citation

Subtleties in the trainability of quantum machine learning models

Supanut Thanasilp, Samson Wang, Nhat A. Nghiem +2 more

Abstract A new paradigm for data science has emerged, with quantum data, quantum models, and quantum computational devices. This field, called quantum machine learning (QML), aims to achieve a speedup over traditional machine learning for data analysis. However, its success usually hinges on efficiently training the parameters in quantum neural networks, and the field of QML is still lacking theoretical scaling results for their trainability. Some trainability results have been proven for a closely related field called variational quantum algorithms (VQAs). While both fields involve training a parametrized quantum circuit, there are crucial differences that make the results for one setting not readily applicable to the other. In this work, we bridge the two frameworks and show that gradient scaling results for VQAs can also be applied to study the gradient scaling of QML models. Our results indicate that features deemed detrimental for VQA trainability can also lead to issues such as barren plateaus in QML. Consequently, our work has implications for several QML proposals in the literature. In addition, we provide theoretical and numerical evidence that QML models exhibit further trainability issues not present in VQAs, arising from the use of a training dataset. We refer to these as dataset-induced barren plateaus. These results are most relevant when dealing with classical data, as here the choice of embedding scheme (i.e., the map between classical data and quantum states) can greatly affect the gradient scaling.

arXiv:2110.14753•Quantum Machine Intelligence•May 15, 2023•115 citations

Software Between Quantum and Machine Learning -- And Down to Pulses

Maja Franz, Melvin Strobl, Jonathan Hunz +5 more

Contemporary quantum computing platforms remain, in essence, programmable physical systems whose control is typically mediated through unitary gate abstractions. While such abstractions provide a uniform interface, they obscure important aspects of the underlying hardware and may limit the exploitation of its full capabilities. Direct operation at the control-pulse level offers a more expressive and physically faithful paradigm, enabling, for instance, the implementation of tailored error-mitigation and optimisation strategies. However, this increased expressivity comes at the cost of greater quantum software development complexity, necessitating structured and accessible tooling. We present a software framework, integrated within the QML-Essentials package, that extends quantum machine learning (QML) methodologies to encompass pulse-level modelling. By embedding quantum optimal control techniques within a QML setting, our approach enables the seamless combination of gate-based and pulse-level representations. The framework provides a comprehensive suite of modelling and analytical capabilities. In particular, we introduce composable ansatz constructions based on interchangeable building blocks, and support for end-to-end optimisation of pulse parameters. Motivated by the central role of quantum Fourier models, we further incorporate a range of Fourier-analytic diagnostics, complemented by extended measures of entanglement. All performance-critical components are implemented in a high-performance environment using JAX and supported by a dedicated quantum simulator. Taken together, the framework facilitates reproducible and systematic investigations, while bridging the conceptual and practical divide between abstract circuit models and hardware-aware optimisation. It provides a robust foundation for future developments at the intersection of QML and quantum control.

quant-pharXiv:2605.21286•IEEE International Conference on Quantum Software (QSW)•May 21, 2026•2 citations
Pulse-based quantum machine learning (QML) has emerged as a promising paradigm in quantum artificial intelligence due to its exceptional hardware efficiency. For real-world applications, it is crucial that pulse-based models are both expressive and trainable. Previous studies have shown that pulse-based models operating under dynamical symmetry can be effectively trained, benefiting from a favorable loss landscape devoid of barren plateaus. However, this approach may lead to uncontrollability, which can limit expressivity if the model is poorly designed. In this paper, we explore the conditions necessary for pulse-based QML models to maintain both expressivity and trainability. We present a key requirement involving the system's initial state, the measurement observable, and the underlying dynamical symmetry Lie algebra, supported by numerical simulations. Our results provide a comprehensive framework for the design of practical pulse-based QML models that strike a good balance between expressivity and trainability.
Pulse-based quantum machine learning (QML) models possess full expressivity when they are ensemble controllable. However, the accompanied barren plateaus render training intractable for high-dimensional systems. In this paper, we show that the trade-off can be understood via controllability analysis. We first apply the Fliess-series expansion to pulse-based QML models to investigate the effect of control system structure on model expressivity, which leads to a universal criterion for assessing the expressivity of generic QML models. We then show that uncontrollable pulse-based models that evolve on low-dimensional manifolds can achieve better balance between expressivity and trainability. Numerical experiments verify the proposed criterion and our analysis, which further demonstrate that the expressivity can be enhanced by increasing dimensionality, while barren plateaus can be avoided by reducing the controllability. Our approach provides a promising path for designing pulse-based QML models that are both highly expressive and trainable.
Recent advances in Machine Learning have transformed numerous industrial sectors, yet classical paradigms face fundamental limitations: rapidly growing data volumes, rising computational costs, significant energy consumption, and the physical scaling limits of conventional hardware architectures. Quantum computing has emerged as a promising computational paradigm to address these challenges, giving rise to the field of Quantum Machine Learning (QML). In this thesis, the theoretical foundations of QML are investigated, with a focus on near-term and future practical applications. Three central challenges are addressed: the trainability of variational quantum circuits, their expressivity, and their resistance to efficient classical simulation. The trainability of Hamming-weight preserving variational quantum circuits is first studied, and theoretical guarantees are established that resolve an open conjecture on the absence of barren plateaus for this circuit family. Subspace-preserving QML algorithms are then introduced, including photonic circuits and quantum convolutional neural networks, and are designed to mimic classical ML subroutines while offering polynomial quantum advantage. Finally, variational quantum circuits are analyzed as quantum Fourier models, and a framework is derived to jointly characterize expressivity and trainability, from which conditions are obtained under which quantum models provably separate from their classical counterparts. These contributions are intended to advance the theoretical roadmap for harnessing near-term and future quantum technologies in real-world applications.
Exploiting the power of quantum computation to realize superior machine learning algorithms has been a major research focus of recent years, but the prospects of quantum machine learning (QML) remain dampened by considerable technical challenges. A particularly significant issue is that generic QML models suffer from so-called barren plateaus in their training landscapes—large regions where cost function gradients vanish exponentially in the number of qubits employed, rendering large models effectively untrainable. A leading strategy for combating this effect is to build problem-specific models that take into account the symmetries of their data in order to focus on a smaller, relevant subset of Hilbert space. In this work, we introduce a family of rotationally equivariant QML models built upon the quantum Fourier transform, and leverage recent insights from the Lie-algebraic study of QML models to prove that (a subset of) our models do not exhibit barren plateaus. In addition to our analytical results we numerically test our rotationally equivariant models on a dataset of simulated scanning tunneling microscope images of phosphorus impurities in silicon, where rotational symmetry naturally arises, and find that they dramatically outperform their generic counterparts in practice. Published by the American Physical Society 2024
As Quantum Machine Learning (QML) transitions toward practical implementation, the field faces a critical architectural bottleneck that challenges the fundamental assumptions of classical statistical learning theory. In classical deep learning, increasing model capacity typically risks overfitting. However, this study advances a counter-intuitive paradigm: unstructured contemporary QML architectures suffer from a profound state of quantum underfitting, driven by the "expressivity-trainability paradox." We demonstrate that the vast Hilbert space capacity of Parameterized Quantum Circuits (PQCs)-traditionally chased as the source of quantum advantage is the direct mathematical cause of Barren Plateaus (BPs), where gradient landscapes become exponentially flat. By synthesizing recent breakthroughs in Dynamical Lie Algebras (DLAs) and Geometric QML, we establish a comprehensive framework linking the algebraic dimension of circuit generators to their optimization dynamics. Furthermore, we empirically validate this framework on a non-linear binary classification task, illuminating a uniquely quantum manifestation of the bias-variance tradeoff: while unstructured architectures achieve near-perfect training accuracy via unscalable parameterization (quantum overfitting), embedding group-theoretic geometric priors acts as a structural regularizer. By restricting the DLA growth to a polynomial regime, our symmetry-preserving approach sacrifices raw memorization capacity to guarantee scalable, gradient-rich training landscapes, offering a robust roadmap for "Trainability-by-Design" in scalable quantum neural networks.
Abstract A new paradigm for data science has emerged, with quantum data, quantum models, and quantum computational devices. This field, called quantum machine learning (QML), aims to achieve a speedup over traditional machine learning for data analysis. However, its success usually hinges on efficiently training the parameters in quantum neural networks, and the field of QML is still lacking theoretical scaling results for their trainability. Some trainability results have been proven for a closely related field called variational quantum algorithms (VQAs). While both fields involve training a parametrized quantum circuit, there are crucial differences that make the results for one setting not readily applicable to the other. In this work, we bridge the two frameworks and show that gradient scaling results for VQAs can also be applied to study the gradient scaling of QML models. Our results indicate that features deemed detrimental for VQA trainability can also lead to issues such as barren plateaus in QML. Consequently, our work has implications for several QML proposals in the literature. In addition, we provide theoretical and numerical evidence that QML models exhibit further trainability issues not present in VQAs, arising from the use of a training dataset. We refer to these as dataset-induced barren plateaus. These results are most relevant when dealing with classical data, as here the choice of embedding scheme (i.e., the map between classical data and quantum states) can greatly affect the gradient scaling.
Contemporary quantum computing platforms remain, in essence, programmable physical systems whose control is typically mediated through unitary gate abstractions. While such abstractions provide a uniform interface, they obscure important aspects of the underlying hardware and may limit the exploitation of its full capabilities. Direct operation at the control-pulse level offers a more expressive and physically faithful paradigm, enabling, for instance, the implementation of tailored error-mitigation and optimisation strategies. However, this increased expressivity comes at the cost of greater quantum software development complexity, necessitating structured and accessible tooling. We present a software framework, integrated within the QML-Essentials package, that extends quantum machine learning (QML) methodologies to encompass pulse-level modelling. By embedding quantum optimal control techniques within a QML setting, our approach enables the seamless combination of gate-based and pulse-level representations. The framework provides a comprehensive suite of modelling and analytical capabilities. In particular, we introduce composable ansatz constructions based on interchangeable building blocks, and support for end-to-end optimisation of pulse parameters. Motivated by the central role of quantum Fourier models, we further incorporate a range of Fourier-analytic diagnostics, complemented by extended measures of entanglement. All performance-critical components are implemented in a high-performance environment using JAX and supported by a dedicated quantum simulator. Taken together, the framework facilitates reproducible and systematic investigations, while bridging the conceptual and practical divide between abstract circuit models and hardware-aware optimisation. It provides a robust foundation for future developments at the intersection of QML and quantum control.