The results demonstrate why controlled component attribution is necessary before crediting a hybrid model's performance to its quantum layer, and analyze bottleneck, simulation, finite-shot, and noise limitations.
Abstract
We test whether a parameterized quantum circuit (PQC) improves a hybrid quantum-classical model's performance on classical datasets, using an interface-matched classical map as the control while holding all other components fixed. Our architecture, Quantum-Embedded Attention (QEA), uses a learnable projector to compress backbone features into an $n_q$-dimensional angle vector, a shallow PQC to map those angles to one- and two-qubit Pauli expectations, and a classical attention decoder to produce class logits. We hypothesized the PQC would improve accuracy or seed-to-seed stability over a classical map with matched input/output dimensions. We test this with an interface-matched $2\times2$ factorial on Breast Cancer Wisconsin at $n_q\in\{4,8\}$, independently swapping the PQC for a classical map and the attention decoder for a linear head, across five paired seeds per cell. Three of four paired quantum-minus-classical $95\%$ confidence intervals include zero; the fourth, a $+1.63$ percentage-point contrast for the attention decoder at $n_q=4$, reverses sign at $n_q=8$ and does not survive correction across the four contrasts. The experiment thus shows no consistent PQC contribution and cannot establish equivalence. A five-dataset cross-modality grid shows comparable accuracy on AG~News, Breast Cancer Wisconsin, and BirdCLEF but a large deficit on CIFAR-10; these cells are not interface-matched and are interpreted descriptively. We report all planned canonical runs, distinguish current Pauli-readout results from legacy probability-readout experiments, and analyze bottleneck, simulation, finite-shot, and noise limitations. The results do not establish a quantum advantage; they demonstrate why controlled component attribution is necessary before crediting a hybrid model's performance to its quantum layer.
Neural quantum kernels (NQKs) construct quantum kernels by pretraining a quantum neural network (QNN) and subsequently reusing the trained circuit as a task-adapted embedding. Extending this framework to qudits, with local unitaries in $\mathrm{SU}(d)$, provides a natural route to richer data embeddings through the increased local degrees of freedom and a direct interface for multiclass classification via intrinsically multi-level quantum systems. In this work, focusing on qutrits ($d=3$), we extend NQKs to the qudit setting and perform a systematic study of key design choices, including the number of encoded features, the number of qutrits, the kernel construction (1-to-$n$ and $n$-to-$n$), and the parameterization of $\mathrm{SU}(3)$ unitaries. Across binary and three-class tasks on four benchmark datasets, qutrit NQKs improve over the corresponding QNN baselines in nearly all settings considered and can benefit from scaling both the feature budget and the system size, although the magnitude of these gains may saturate, is dataset-dependent, and depends on the chosen parameterization. In particular, an ablation over $\mathrm{SU}(3)$ parameterizations shows that the unitary representation can substantially impact both optimization behaviour and classifier performance. These findings highlight the potential of qudit-based quantum models not only as a straightforward generalization of qubit-based architectures, but also as a promising means to better exploit complex data structures in quantum machine learning.
Camila Cristiano-Romero, Pablo Rodriguez-Grasa, Mikel Sanz· 0 citations
The way in which classical data are encoded into quantum states plays a significant role in both classification performance and quantum circuit complexity in Quantum Machine Learning. In this study, the effects of different quantum feature encoding strategies on Quantum Support Vector Machine performance were investigated using five binary classification datasets. In particular, the statistical relationships between features were incorporated into quantum circuits through \(RY(\theta)\) and controlled-\(RY(\theta)\) gates, and this approach was compared with conventional quantum feature maps. The results demonstrate that incorporating statistical relationships into the encoding process can influence classification performance. However, more complex and densely entangled circuits do not necessarily yield higher performance. In addition, a composite evaluation metric was employed to jointly assess predictive performance, generalization, and circuit cost. The findings across the five datasets indicate that the choice of quantum feature encoding strategy should account for the underlying structure of the data and that predictive performance should be evaluated together with quantum circuit complexity.
The intersection of quantum processing and classical machine learning has spawned hybrid quantum-classical systems — practical system designs that attempt to use quantum potential in the limitations of modern Noisy Intermediate-Scale Quantum (NISQ) hardware. This paper introduces the development, deployment, and empirical analysis of a hybrid quantum-classical image classifier (which combines a Convolutional Neural Network (CNN) with an eight-qubit Variational Quantum Circuit (VQC)) in the binary classification of handwritten digits. This implementation, implemented in PennyLane and PyTorch, reaches a peak test accuracy of 99.85% across 1,984 test samples of the MNIST system, and only three errors are made. In addition to performance measures, the work presents an approach based on Quantum Software Engineering (QSE) by reporting major engineering issues, such as quantum-classical interface design, adjoint differentiation, feature-dimensionality reduction, and backend portability, and suggestion of seven quality-assurance practices of hybrid quantum software systems. The results demonstrate that it is possible to manufacture successfully hybrid quantum-classical architectures with the help of existing open-source tools and simulators, and provide future research and practice with hybrid QSE with effective advice.
Haider Ali, Muhammad Azeem Akbar, A. Khan et al.· SIGSOFT FSE Companion· 0 citations
Hybrid classical-quantum machine learning is a promising approach for image classification, but current methods often require too many qubits (quantum resources), creating a bottleneck for practical use. Current methodologies either constrain classification to basic binary decisions due to intricate quantum circuit design or depend on a Convolutional Neural Network (CNN) feature extractor that interfaces with a quantum layer necessitating a substantial quantity of qubits, frequently equivalent to the number of extracted features. To solve this, we propose a new Hybrid CNN-Quantum framework that dramatically reduces the required quantum resources. Our key innovation is an amplitude-encoding-inspired technique and a new activation function that together allow us to infer the final classification using only ⌈log2 (number of classes)⌉ readout qubits. We validated our framework on MNIST, Fashion-MNIST, and KMNIST, achieving accuracies of 0.9697, 0.8515, and 0.9258, respectively, using just 8 qubits. This matches or surpasses prior results that used over 1500 qubits, highlighting our method's competitive accuracy with drastically reduced quantum hardware requirements.
The exponential parameter scaling of classical transformer models confronts severe physical and economic barriers. To sustain generative AI capabilities, alternative computational paradigms must be explored.
This paper projects the architecture and scaling laws of Quantum Generative AI Foundation Models by integrating Variational Quantum Algorithms (VQAs) with Fault-Tolerant Quantum Error Correction (QEC). We propose a hybrid quantum-classical framework utilizing isometric Tree Tensor Networks (TTNs) and a novel Quantum Self-Attention (QSA) subroutine, capable of compressing the latent space of a classical 10
11
-parameter Large Language Model (LLM) into a 10
6
parameter quantum neural network via amplitude encoding.
To circumvent Noise-Induced Barren Plateaus (NIBP), we map the training requirements onto a fault-tolerant regime. Assuming a surface code QEC overhead with a physical-to-logical qubit ratio of approximately 2,000:1, we establish the resource requirements for a VQA operating below the 10
−4
physical gate error threshold. Our numerical projections indicate an approximate 45% reduction in total energy expenditure for frontier model training and a per-query attention processing complexity of
O
(
L
log
d
)
.
The quantum framework fundamentally subverts the classical compute wall by substituting linear parameter scaling with logarithmic latent space compression, acknowledging that full attention matrix computation retains a dependency on measurement precision overheads.
R. Delhibabu· Frontiers of Computer Scienc...· 0 citations
Next-generation HQCNN architectures, including DMERA, HEA–QCNN, and (bQCNN) designs, significantly optimized for NISQ hardware are presented, establishing a scalable, parameter-efficient framework for high-performance hybrid quantum-classical ML on near-term devices.
Shyam R. Sihare, A. Cherukuri· SN Computer Science· 0 citations
