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:2607.24014

Scalable Quantum Machine Learning: Trainability, Expressivity and Efficiency

Iordanis Kerenidis

quant-phSubmitted Jul 27, 2026• Announced Jul 28, 2026• 7 citations• Revised Oct 27, 2026 (v3)
v1 · Jul 27, 2026v2 · Aug 20, 2026v3 · Oct 27, 2026
arXivHTMLPDF

Abstract

Scalable parameterized quantum circuits for machine learning face three obstacles: barren plateaus, no guarantee the learned function class is classically hard, and prohibitive circuit evaluations per gradient step. We propose the unitary brick-wall: a $k$-particle fermionic architecture for nearest-neighbor hardware, combining Reconfigurable Beam Splitters with interleaved single-qubit phases and a non-Gaussian magic-state encoding, where $k$ trades classical hardness against training cost. Trainable. The brick-wall has dynamical Lie algebra $\mathfrak{u}(n)$ and is surjective onto $U(n)$ via Givens rotations, enabling Haar initialization. Two-body correlator readouts achieve gradient variance $Θ(k^3/n^5)$, polynomial in $n$ throughout $n-2k=Ω(n)$. Expressive. Hardness is controlled by $k$: best-known classical sampling runs in time $2^{Θ(k)}$, worst-case #P-hardness holds from $k=n^ε$, and the average-case machinery of Fermion Sampling applies at $k=Θ(n)$. The sampler is input-dependent, so costs are paid per query: at our operating point $k=100$ one query needs $\sim 10^{15}$ operations, beyond commercial query rates, and at $k=120$ an exascale machine sustains under one query per second. Efficient. A multi-layer parallel parameter-shift rule computes all $O(n^2)$ gradients from $4kn$ circuit evaluations per step, a factor $n/k$ reduction over the standard rule's $4n^2$, growing linearly with $n$ at fixed $k$. The unitary butterfly targets all-to-all hardware, with $2\log n$ layers of parallel gates and $(3/2)n\log n$ parameters, similar hardness guarantees, and $4k\log n$ evaluations per step -- the same factor-$n/k$ reduction. Its trainability holds at two levels: absence of exponential barren plateaus is unconditional, while the sharp $Θ(k^3/n^5)$ rate needs a conjectured scalar comparison with the two-particle Haar second moment.

Comments
47 pages. v3: Fermionic beam splitter gates on non-adjacent modes. Hadamard spreading circuits corrected. Butterfly gradient-variance lower bound sharpened. Classical sampling cost and operating point updated to the best currently known algorithm. The two-particle design conjecture is restated as a scalar comparison. No trainability, parameter-shift or hardness conclusion changes

Related papers

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

Comment on "Scalable Quantum Machine Learning: Trainability, Expressivity and Efficiency": Polynomial Evaluation of the Triplet-Block Readout

Erfan Amidi

We examine the classical-cost claim for the triplet-block two-body readout in arXiv:2607.24014v1. The Gaussian-state expansion used there gives an $O(2^{2k/3}\mathrm{poly}(n))$ classical algorithm, but it is not necessary for fixed-body observables. The triplet-block input has an explicitly computable diagonal two-particle reduced density matrix, which passive fermionic linear optics propagates through $\bigwedge^2 W$. This gives a deterministic $O(n^4)$ algorithm for the complete correlator vector $(\langle n_i n_j\rangle)_{i<j}$, independently of $k$ and the fermionic-linear-optics extent. More generally, every number-conserving fixed-$r$-body expectation is polynomially computable whenever the input $r$-particle reduced density matrix is classically available; if that matrix is diagonal, all diagonal correlators are computable in $O(n^{2r})$ time. This invalidates the algorithm-relative exponential-cost conclusion for the supervised two-body readout, without affecting the gradient-variance, barren-plateau, parameter-shift, or sampling-hardness results.

quant-pharXiv:2608.20435•Aug 21, 2026•0 citations

Train on classical, deploy on quantum: scaling generative quantum machine learning to a thousand qubits

Erik Recio-Armengol, Shahnawaz Ahmed, Bowles, Joseph

We propose an approach to generative quantum machine learning that overcomes the fundamental scaling issues of variational quantum circuits. The core idea is to use a class of generative models based on instantaneous quantum polynomial circuits, which we show can be trained efficiently on classical hardware. Although training is classically efficient, sampling from these circuits is widely believed to be classically hard, and so computational advantages are possible when sampling from the trained model on quantum hardware. By combining our approach with a data-dependent parameter initialisation strategy, we do not encounter issues of barren plateaus and successfully circumvent the poor scaling of gradient estimation that plagues traditional approaches to quantum circuit optimisation. We investigate and evaluate our approach on a number of real and synthetic datasets, training models with up to one thousand qubits and hundreds of thousands of parameters. We find that the quantum models can successfully learn from high dimensional data, and perform surprisingly well compared to simple energy-based classical generative models trained with a similar amount of hyperparameter optimisation. Overall, our work demonstrates that a path to scalable quantum generative machine learning exists and can be investigated today at large scales.

arXiv:2503.02934•Mar 4, 2025•61 citations

How to find expressible and trainable parameterized quantum circuits?

Peter Röseler, Dennis Willsch, Kristel Michielsen

Whether parameterized quantum circuits (PQCs) can be systematically constructed to be both trainable and expressive remains an open question. Highly expressive PQCs often exhibit barren plateaus, while several trainable alternatives admit efficient classical simulation. We address this question by deriving a finite-sample, dimension-independent concentration bound for estimating the variance of a PQC cost function, yielding explicit trainability guarantees. Across commonly used ansätze, we observe an anticorrelation between trainability and expressibility, consistent with theoretical insights. Building on this observation, we propose a property-based ansatz-search framework for identifying circuits that combine trainability and expressibility. We demonstrate its practical viability on a real quantum computer and apply it to variational quantum algorithms. We identify quantum neural network ansätze with improved effective dimension using over $6 \times$ fewer parameters, and for VQE on $\mathrm{H}_2$ we achieve UCCSD-like accuracy at substantially reduced circuit complexity.

quant-pharXiv:2603.14451•Mar 17, 2026•7 citations

Special-Unitary Parameterization for Trainable Variational Quantum Circuits

Kuan-Cheng Chen, Hua-an Tseng, Samuel Yen-Chi Chen +2 more

We propose SUN-VQC, a variational-circuit architecture whose elementary layers are single exponentials of a symmetry-restricted Lie subgroup, $\mathrm{SU}(2^{k}) \subset \mathrm{SU}(2^{n})$ with $k \ll n$. Confining the evolution to this compact subspace reduces the dynamical Lie-algebra dimension from $\mathcal{O}(4^{n})$ to $\mathcal{O}(4^{k})$, ensuring only polynomial suppression of gradient variance and circumventing barren plateaus that plague hardware-efficient ansätze. Exact, hardware-compatible gradients are obtained using a generalized parameter-shift rule, avoiding ancillary qubits and finite-difference bias. Numerical experiments on quantum auto-encoding and classification show that SUN-VQCs sustain order-of-magnitude larger gradient signals, converge 2--3$\times$ faster, and reach higher final fidelities than depth-matched Pauli-rotation or hardware-efficient circuits. These results demonstrate that Lie-subalgebra engineering provides a principled, scalable route to barren-plateau-resilient VQAs compatible with near-term quantum processors.

arXiv:2507.05535•IEEE International Conference on Quantum Artificial Intelligence (QAI)•Nov 2, 2025•1 citation

Quantum Machine Learning (QML) and Parameterized Quantum Circuits for High-Dimensional Optimization: Barren Plateaus, Quantum Kernels, and NISQ Benchmarks

Harshit Tinker, Divyajeet Swami, Vishal Shrivastava +3 more

Abstract—Quantum Machine Learning (QML) operating on Noisy Intermediate-Scale Quantum (NISQ) hardware represents an emerging computational paradigm to address high-dimensional optimization problems that challenge classical Von Neumann supercomputing architectures. By mapping complex feature vectors into exponentially large 2^N Hilbert state spaces via Parameterized Quantum Circuits (PQCs) and Variational Quantum Algorithms (VQAs), QML targets computational advantages in combinatorial optimization, high-dimensional classification, and quantum-enhanced molecular modeling. However, the theoretical promise of NISQ-era variational computing is severely constrained by three fundamental bottlenecks: (1) Barren Plateaus, characterized by the exponential concentration of measure and vanishing variance of cost function gradients; (2) Hardware Decoherece Noise and readout error rates that corrupt quantum state fidelity; and (3) Classical-to-Quantum State Preparation bottlenecks that can erase theoretical computational speedups. This paper delivers an exhaustive, mathematically formal technical investigation and empirical benchmark of QML architectures for high-dimensional optimization. We formalize the Haar-distributed geometric mechanics of barren plateaus under 2-design unitary assumptions, evaluate barren plateau mitigation through shallow local cost operators and layer-by-layer parameter initialization, and contrast Variational Quantum Eigensolvers (VQE) with Quantum Kernel Support Vector Machines (Q-SVM). Furthermore, we construct an empirical benchmarking testbed executing across both state-vector simulators and superconducting quantum backends (IBM Quantum Heron / Eagle architectures), evaluating 127-qubit Hamiltonian optimization and high-dimensional financial portfolio allocation. Our findings demonstrate that while global cost functions suffer exponential gradient decay Var[grad] ~ O(2^-N) across circuit depth d >= 4, local Hamiltonian decomposition maintains trainable variance, achieving a 98.2% ground-state fidelity under zero-noise extrapolation (ZNE) error mitigation. Finally, we formulate open challenges in quantum error mitigation scaling, non-perturbative barren plateau escapes, and quantum data loading efficiency. Index Terms—Quantum Machine Learning (QML), Parameterized Quantum Circuits (PQCs), Variational Quantum Eigensolver (VQE), Barren Plateaus, Quantum Kernel Methods, NISQ Computing, Error Mitigation, Hilbert Space.

•Oct 13, 2026•0 citations
Scalable parameterized quantum circuits for machine learning face three obstacles: barren plateaus, no guarantee the learned function class is classically hard, and prohibitive circuit evaluations per gradient step. We propose the unitary brick-wall: a kkk-particle fermionic architecture for nearest-neighbor hardware, combining Reconfigurable Beam Splitters with interleaved single-qubit phases and a non-Gaussian magic-state encoding, where kkk trades classical hardness against training cost.
Trainable. The brick-wall has dynamical Lie algebra u(n)\mathfrak{u}(n)u(n) and is surjective onto U(n)U(n)U(n) via Givens rotations, enabling Haar initialization. Two-body correlator readouts achieve gradient variance Θ(k3/n5)Θ(k^3/n^5)Θ(k3/n5), polynomial in nnn throughout n−2k=Ω(n)n-2k=Ω(n)n−2k=Ω(n).
Expressive. Hardness is controlled by kkk: best-known classical sampling runs in time 2Θ(k)2^{Θ(k)}2Θ(k), worst-case #P-hardness holds from k=nεk=n^εk=nε, and the average-case machinery of Fermion Sampling applies at k=Θ(n)k=Θ(n)k=Θ(n). The sampler is input-dependent, so costs are paid per query: at our operating point k=100k=100k=100 one query needs ∼1015\sim 10^{15}∼1015 operations, beyond commercial query rates, and at k=120k=120k=120 an exascale machine sustains under one query per second.
Efficient. A multi-layer parallel parameter-shift rule computes all O(n2)O(n^2)O(n2) gradients from 4kn4kn4kn circuit evaluations per step, a factor n/kn/kn/k reduction over the standard rule's 4n24n^24n2, growing linearly with nnn at fixed kkk.
The unitary butterfly targets all-to-all hardware, with 2log⁡n2\log n2logn layers of parallel gates and (3/2)nlog⁡n(3/2)n\log n(3/2)nlogn parameters, similar hardness guarantees, and 4klog⁡n4k\log n4klogn evaluations per step -- the same factor-n/kn/kn/k reduction. Its trainability holds at two levels: absence of exponential barren plateaus is unconditional, while the sharp Θ(k3/n5)Θ(k^3/n^5)Θ(k3/n5) rate needs a conjectured scalar comparison with the two-particle Haar second moment.
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.
We examine the classical-cost claim for the triplet-block two-body readout in arXiv:2607.24014v1. The Gaussian-state expansion used there gives an O(22k/3poly(n))O(2^{2k/3}\mathrm{poly}(n))O(22k/3poly(n)) classical algorithm, but it is not necessary for fixed-body observables. The triplet-block input has an explicitly computable diagonal two-particle reduced density matrix, which passive fermionic linear optics propagates through ⋀2W\bigwedge^2 W⋀2W. This gives a deterministic O(n4)O(n^4)O(n4) algorithm for the complete correlator vector (⟨ninj⟩)i<j(\langle n_i n_j\rangle)_{i<j}(⟨ni​nj​⟩)i<j​, independently of kkk and the fermionic-linear-optics extent. More generally, every number-conserving fixed-rrr-body expectation is polynomially computable whenever the input rrr-particle reduced density matrix is classically available; if that matrix is diagonal, all diagonal correlators are computable in O(n2r)O(n^{2r})O(n2r) time. This invalidates the algorithm-relative exponential-cost conclusion for the supervised two-body readout, without affecting the gradient-variance, barren-plateau, parameter-shift, or sampling-hardness results.
We propose an approach to generative quantum machine learning that overcomes the fundamental scaling issues of variational quantum circuits. The core idea is to use a class of generative models based on instantaneous quantum polynomial circuits, which we show can be trained efficiently on classical hardware. Although training is classically efficient, sampling from these circuits is widely believed to be classically hard, and so computational advantages are possible when sampling from the trained model on quantum hardware. By combining our approach with a data-dependent parameter initialisation strategy, we do not encounter issues of barren plateaus and successfully circumvent the poor scaling of gradient estimation that plagues traditional approaches to quantum circuit optimisation. We investigate and evaluate our approach on a number of real and synthetic datasets, training models with up to one thousand qubits and hundreds of thousands of parameters. We find that the quantum models can successfully learn from high dimensional data, and perform surprisingly well compared to simple energy-based classical generative models trained with a similar amount of hyperparameter optimisation. Overall, our work demonstrates that a path to scalable quantum generative machine learning exists and can be investigated today at large scales.
Whether parameterized quantum circuits (PQCs) can be systematically constructed to be both trainable and expressive remains an open question. Highly expressive PQCs often exhibit barren plateaus, while several trainable alternatives admit efficient classical simulation. We address this question by deriving a finite-sample, dimension-independent concentration bound for estimating the variance of a PQC cost function, yielding explicit trainability guarantees. Across commonly used ansätze, we observe an anticorrelation between trainability and expressibility, consistent with theoretical insights. Building on this observation, we propose a property-based ansatz-search framework for identifying circuits that combine trainability and expressibility. We demonstrate its practical viability on a real quantum computer and apply it to variational quantum algorithms. We identify quantum neural network ansätze with improved effective dimension using over 6×6 \times6× fewer parameters, and for VQE on H2\mathrm{H}_2H2​ we achieve UCCSD-like accuracy at substantially reduced circuit complexity.
We propose SUN-VQC, a variational-circuit architecture whose elementary layers are single exponentials of a symmetry-restricted Lie subgroup, SU(2k)⊂SU(2n)\mathrm{SU}(2^{k}) \subset \mathrm{SU}(2^{n})SU(2k)⊂SU(2n) with k≪nk \ll nk≪n. Confining the evolution to this compact subspace reduces the dynamical Lie-algebra dimension from O(4n)\mathcal{O}(4^{n})O(4n) to O(4k)\mathcal{O}(4^{k})O(4k), ensuring only polynomial suppression of gradient variance and circumventing barren plateaus that plague hardware-efficient ansätze. Exact, hardware-compatible gradients are obtained using a generalized parameter-shift rule, avoiding ancillary qubits and finite-difference bias. Numerical experiments on quantum auto-encoding and classification show that SUN-VQCs sustain order-of-magnitude larger gradient signals, converge 2--3×\times× faster, and reach higher final fidelities than depth-matched Pauli-rotation or hardware-efficient circuits. These results demonstrate that Lie-subalgebra engineering provides a principled, scalable route to barren-plateau-resilient VQAs compatible with near-term quantum processors.
Abstract—Quantum Machine Learning (QML) operating on Noisy Intermediate-Scale Quantum (NISQ) hardware represents an emerging computational paradigm to address high-dimensional optimization problems that challenge classical Von Neumann supercomputing architectures. By mapping complex feature vectors into exponentially large 2^N Hilbert state spaces via Parameterized Quantum Circuits (PQCs) and Variational Quantum Algorithms (VQAs), QML targets computational advantages in combinatorial optimization, high-dimensional classification, and quantum-enhanced molecular modeling. However, the theoretical promise of NISQ-era variational computing is severely constrained by three fundamental bottlenecks: (1) Barren Plateaus, characterized by the exponential concentration of measure and vanishing variance of cost function gradients; (2) Hardware Decoherece Noise and readout error rates that corrupt quantum state fidelity; and (3) Classical-to-Quantum State Preparation bottlenecks that can erase theoretical computational speedups. This paper delivers an exhaustive, mathematically formal technical investigation and empirical benchmark of QML architectures for high-dimensional optimization. We formalize the Haar-distributed geometric mechanics of barren plateaus under 2-design unitary assumptions, evaluate barren plateau mitigation through shallow local cost operators and layer-by-layer parameter initialization, and contrast Variational Quantum Eigensolvers (VQE) with Quantum Kernel Support Vector Machines (Q-SVM). Furthermore, we construct an empirical benchmarking testbed executing across both state-vector simulators and superconducting quantum backends (IBM Quantum Heron / Eagle architectures), evaluating 127-qubit Hamiltonian optimization and high-dimensional financial portfolio allocation. Our findings demonstrate that while global cost functions suffer exponential gradient decay Var[grad] ~ O(2^-N) across circuit depth d >= 4, local Hamiltonian decomposition maintains trainable variance, achieving a 98.2% ground-state fidelity under zero-noise extrapolation (ZNE) error mitigation. Finally, we formulate open challenges in quantum error mitigation scaling, non-perturbative barren plateau escapes, and quantum data loading efficiency. Index Terms—Quantum Machine Learning (QML), Parameterized Quantum Circuits (PQCs), Variational Quantum Eigensolver (VQE), Barren Plateaus, Quantum Kernel Methods, NISQ Computing, Error Mitigation, Hilbert Space.