Auto Quantum Machine Learning for Multisource Classification
Authors: Tomasz Rybotycki, Sebastian Dziura, Piotr Gawron
Organizations: Center of Excellence in Artificial Intelligence, AGH University, al. Mickiewicza 30, 30-059 Cracow, Poland · Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences, ul. Bartycka 18, 00-716 Warsaw, Poland · Systems Research Institute, Polish Academy of Sciences, ul. Newelska 6, 01-447 Warsaw, Poland
With fault-tolerant quantum computing on the horizon, there is growing interest in applying quantum computational methods to data-intensive scientific fields like remote sensing. Quantum machine learning (QML) has already demonstrated potential for such demanding tasks. One area of particular focus is quantum data fusion -- a complex data analysis problem that has attracted significant recent attention. In this work, we introduce an automated QML (AQML) approach for addressing data fusion challenges. We evaluate how AQML-generated quantum circuits perform compared to classical multilayer perceptrons (MLPs) and manually designed QML models when processing multisource inputs. Furthermore, we apply our method to change detection using the multispectral ONERA dataset, achieving improved accuracy over previously reported QML-based change detection results.
Figures & tables
Figure 1: A schematic representation of all the models used in this work: yellow — feature extractor, blue — classifier (here the fusion takes place), green — optional output logits for quantum classifiers. XOR on the schematic means that only one type of classifier is used at the time. Top classifier represents MLP approach, and the bottom one parametrized quantum circuit approach.
Figure 2: A general schematic of a parameterized quantum circuit. It consists of K QML blocks, each containing a data (re-)uploading operation and a variational operation. Notice that both Uload and Uvar could, in principle, be an identity. On the schematic, x denotes the input, and θ denotes the full set of PQC parameters. The PQC is concluded with a measurement.
Figure 3: Overview of the synthetic multisource preprocessing pipeline. (a) Original input image. (b) Horizontal splitting into top and bottom halves. (c) Binarization of each half. (d) Spatial downsampling. (e) Computation of column-wise mean pixel intensities, resulting in vector representations that constitute the synthetic multisource inputs xtop and xbottom .
Parameter
Range
Type
Best value
Learning rate
[10−3,10−2]
linear
10−3
Batch size
[32,128]
discrete
83
Hidden layer size (feature extractors)
[64,256]
discrete
90
Output layer size (feature extractors)
[64,256]
discrete
196
Hidden layer size (classifier)
[64,256]
discrete
95
Table 1: Experiment setup for baseline hyperparameters optimization.
Parameter
Range
Type
Best value
PQC Manual ( AngleX →3× BEL )
Batch size
[32,128]
discrete
32
Hidden layer size (feature extractors)
[64,256]
discrete
128
Output layer size (feature extractors)
[4,12]
discrete
6
BasicEntangler layers number (classifier)
[1,5]
discrete
3
PQC AQMLSolo ( Amplitude → SEL → AngleY →2× SEL )
Table 2: Experiment setup for PQC hyperparameters optimization. For those experiments the learning rate was fixed to lr=10−3 . PQC AQMLSolo and PQC AQMLLinear denote PQC models without and with final classical linear layer respectively. We present the architecture of the model in the parenthesis, next to the model’s name.
Table 3: Average accuracy and f1 macro -score metrics for classical, manual PQC and AQML-found PQC models. PQC AQMLSolo and PQC AQMLLinear denote PQC models without and with final classical linear layer respectively. We also provide trainable parameters number for each model.
Figure 4: A pair of images from the ONERA dataset. They show Saclay, a city in France. The images were taken on 15 III 2016 (Left) and 29 VIII 2017 (Right).
Recent studies have reported promising results from quantum machine learning in drug discovery, healthcare, materials science and cybersecurity. Research includes comparisons with classical machine learning algorithms and the development of sequential hybrid pipelines. We propose that fusion can improve prediction when classical and quantum representations capture complementary predictive relationships. Our classical-quantum multi-view framework preserves the full classical input alongside simulated quantum features. Concatenation and attention combine the views; single-view comparisons, replacement and masking controls assess both views' contributions. Controlled angular-polynomial experiments distinguish targets adequately represented by a single view from those that depend on complementary angular and polynomial relationships. For these mixed targets, fusion improves prediction by combining the relevant feature sets, while matched classical representations recover the same benefit, linking the improvement to their complementary predictive content. Controlled classification tests further show that gains depend on alignment between the chosen representation and the target, and on a suitable fusion rule. Benchmark validation shows improved mean predictive performance over classical baselines on Wine, Wisconsin Diagnostic Breast Cancer, three-class Fashion-MNIST and Steel Plates Faults within the evaluated protocols. Classical replacement and masking controls remain competitive, classical baselines lead on Digits and Covertype, and outcomes with missing or corrupted views depend on the available inputs. These findings support evaluating quantum representations for their complementary predictive contribution and matching the representation and fusion strategy to the target relationships.
Circuit cutting lets a large quantum neural network (QNN) run as independent subcircuits on small devices, but rebuilding its outputs by reconstruction carries a classical sampling overhead exponential in the number of cuts - the dominant runtime cost in prior work. We ask whether, for machine-learning tasks, this step is necessary, and replace it with late fusion: each subcircuit is trained and measured independently, and a small classical head combines their outputs - a linear-cost, decision-level combination borrowed from multimodal learning. To characterize the trade-off we introduce a quantumness dial Q, a tunable reconstruction budget interpolating from pure fusion to full reconstruction, and a cut-entanglement diagnostic that indicates how much reconstruction a task needs (Spearman ρ=0.59 over 104 runs). Across synthetic and standard datasets, independently trained late fusion matches full reconstruction accuracy within 0.04 at every point of the controlled sweep and on every classical benchmark, at exponentially lower cost; it is also markedly more robust to shot and device noise. Controlled entangled-data experiments locate the boundary where fusion must fail. We do not claim advantage over classical machine learning - consistent with recent benchmarking, quantum offers no accuracy edge on these datasets. Late fusion is thus an efficient, noise-robust, self-characterizing alternative to reconstruction for circuit-cutting QML.
Prabhjot Singh, Adel N. Toosi, Rajkumar Buyya
Quantum Cloud Computing and Distributed Systems (qCLOUDS) Lab, School of Computing and Information Systems, The University of Melbourne, Melbourne, VIC, 3010, Australia
A potential path forward is Quantum Machine Learning (QML), which aims to leverage quantum computing in conjunction with classical machine learning to enhance computing efficiency and the expressiveness of models. In this paper, two different quantum classifiers - Variational Quantum Classifier (VQC) and Quantum Kernel Support Vector Machine (QSVM) - are compared with three classical classifiers as baseline classifiers - Logistic Regression, Support Vector Machine (SVM), and a Multi-Layer Perceptron (MLP) - on the Breast Cancer Wisconsin dataset. The quantum circuits were created in the PennyLane framework and simulated on a classical backend. However, in terms of accuracy, classical Logistic Regression performed better with an accuracy of 97.8%, classical SVM and QSVM with an accuracy of 95.6% each, although the Quantum VQC achieved a lower accuracy of 88.9% and had a recall of 100% for the benign class, though it correctly identified only 12 of the 17 malignant cases (a malignant-class recall of approximately 70.6%). The drawback of quantum models is the higher training time; however, since the quantum circuit needs to be classically simulated, the quantum SVM took 23.29 seconds compared to less than 0.01 seconds for the classical linear models. These results indicate that for small structured datasets, classifiers based on quantum computing have not yet surpassed well-tuned classical counterparts. In some respects (e.g., benign-class recall), they perform competitively, though not on malignant-class recall, where the VQC in particular performed worse than the classical baselines, which is worth further investigation on real quantum computers.
Anand Kumar Mishra, Ramanuj Awasthi
School of Engineering & Technology (UIET), CSJM University, Kanpur, Uttar Pradesh, India