Domain-Aware Hybrid Quantum Learning via Correlation-Guided Circuit Design for Crime Pattern Analytics

Niloy Das; Apurba Adhikary; Sheikh Salman Hassan; Tanvir Zaman Khan; Yu Qiao; Zhu Han; Choong Seon Hong

cs.LGApr 8, 2026

Domain-Aware Hybrid Quantum Learning via Correlation-Guided Circuit Design for Crime Pattern Analytics

Authors: Niloy Das, Apurba Adhikary, Sheikh Salman Hassan, Tanvir Zaman Khan, Yu Qiao, Zhu Han, Choong Seon Hong

Organizations: Department of Information and Communication Engineering, Noakhali Science and Technology University, Bangladesh · Institute for Imaging, Data, and Communications (IDCOM), The University of Edinburgh, UK · School of Information Science and Engineering, Zhejiang Sci-Tech University, Hangzhou 310018, China · Department of Electrical and Computer Engineering, University of Houston, Houston, TX 77004-4005, USA ·

Abstract

Crime pattern analysis is critical for law enforcement and predictive policing, yet the surge in criminal activities from rapid urbanization creates high-dimensional, imbalanced datasets that challenge traditional classification methods. This study presents a quantum-classical comparison framework for crime analytics, evaluating four computational paradigms: quantum models, classical baseline machine learning models, and two hybrid quantum-classical architectures. Using 16-year crime statistics, we systematically assess classification performance and computational efficiency under rigorous cross-validation methods. Experimental results show that quantum-inspired approaches, particularly QAOA, achieve up to 84.6% accuracy, while requiring fewer trainable parameters than classical baselines, suggesting practical advantages for memory-constrained edge deployment. The proposed correlation-aware circuit design demonstrates the potential of incorporating domain-specific feature relationships into quantum models. Furthermore, hybrid approaches exhibit competitive training efficiency, making them suitable candidates for resource-constrained environments. The framework's low computational overhead and compact parameter footprint suggest potential advantages for wireless sensor network deployments in smart city surveillance systems, where distributed nodes perform localized crime analytics with minimal communication costs. Our findings provide a preliminary empirical assessment of quantum-enhanced machine learning for structured crime data and motivate further investigation with larger datasets and realistic quantum hardware considerations.

Figures & tables

Fig. 1 : System architecture of the proposed framework showing the complete pipeline.

Fig. 2 : VQC circuit structure (4-qubit, 2-layer). Layer 0 employs correlation-aware CNOT entanglement (weighted by detected feature correlations $\rho>0.5$ ), while Layer 1 uses a standard CNOT ladder for uniform entanglement.

Category	Model	Key Parameters
Quantum (Simulation) †	VQC	4-6 qubits, 2-3 layers
	QAOA	4-6 qubits, 2-3 $p$ -layers
	Q-Kernel SVM	4 qubits, RBF kernel
Q $\rightarrow$ C Hybrid	Q $\rightarrow$ RF	6 qubits $\rightarrow$ 100 trees
	Q $\rightarrow$ SVM	6 qubits $\rightarrow$ RBF kernel
	Q $\rightarrow$ LogReg	6 qubits $\rightarrow$ logistic

TABLE I : Model Configurations

Category	Metric	Definition / Formula
Classification	Accuracy	$\text{Acc}=\frac{\text{TP}+\text{TN}}{N}$ (overall correctness)
	Precision	$\text{Prec}=\frac{\text{TP}}{\text{TP}+\text{FP}}$ (weighted avg. across classes)
	Recall	$\text{Rec}=\frac{\text{TP}}{\text{TP}+\text{FN}}$ (weighted avg. across classes)
	F1-Score	$F_{1}=2\cdot\frac{\text{Prec}\cdot\text{Rec}}{\text{Prec}+\text{Rec}}$ (harmonic mean, weighted)
Quantum Efficiency	Circuit Depth	Total gate count: $d=\sum_{\ell}\|U_{\text{rot}}^{(\ell)}\|+\|U_{\text{ent}}\|$
	Parameter Count	Trainable parameters: $\|\theta\|=n_{q}\times L$ (qubits $\times$ layers)

TABLE II : Evaluation Metrics and Definitions

Fig. 3 : Per-class accuracy comparison highlights quantum advantage on imbalanced categories in preliminary evaluation.

Fig. 4 : Circuit expressibility as a function of quantum layers. Expressibility increases from 0.35 to 0.79 with layer depth.

Quantum-Inspired (Simulation) †
Model	Type	Acc.	Prec.	Rec.	F1
QAOA (4q, 2L)	Quantum	0.85	0.88	0.85	0.85
QAOA (6q, 3L)	Quantum	0.55	0.65	0.55	0.56
VQC (6q, 3L)	Quantum	0.55	0.53	0.55	0.53
VQC (4q, 2L)	Quantum	0.55	0.52	0.55	0.52
QKernel SVM	Quantum	0.40	0.45	0.40	0.42

TABLE III : Preliminary performance comparison (single train–test split). Best result per category in bold .

Model	Type	Acc. $\pm$ CI	F1 $\pm$ CI
Quantum-Inspired (Simulation) †
QAOA (6q, 3L)	Quantum	0.846 $\pm$ 0.019	0.830 $\pm$ 0.021
QAOA (4q, 2L)	Quantum	0.803 $\pm$ 0.024	0.779 $\pm$ 0.026
C $\rightarrow$ Q Hybrid
PCA $\rightarrow$ QAOA	C $\rightarrow$ Q	0.803 $\pm$ 0.024	0.779 $\pm$ 0.026
Best Classical Baseline

TABLE IV : Robust evaluation of top-performing models under 5-fold stratified CV. Best result per category in bold .

Quantum-Inspired
Model	Mean (s)	Params	Infer. $O(\cdot)$	Acc. $\pm$ CI	Sig.
QAOA (4q, 2L)	0.006	16	$O(p\cdot n_{q})$	$0.803\pm 0.024$	∗∗∗
QAOA (6q, 3L)	0.004	36	$O(p\cdot n_{q})$	$0.846\pm 0.019$	∗∗∗
VQC (4q, 2L)	0.015	8	$O(L\cdot n_{q})$	$0.676\pm 0.022$	∗∗∗
VQC (6q, 3L)	0.027	18	$O(L\cdot n_{q})$	$0.686\pm 0.021$	∗∗∗
QKernel SVM	0.075	—	$O(N\cdot d)$	$0.359\pm 0.028$	∗∗∗

TABLE V : Training efficiency, theoretical complexity, and statistical significance.

Quantum Machine Learning

Quantum machine learning uses quantum circuits and quantum data to build learning algorithms and study their computational and statistical limits.

See all 256 papers

Federating Quantum and Classical Computing: A Privacy-Preserving Hybrid Approach
Sep 18, 2026Carlos Cano, Daniel M. Jimenez-Gutierrez, Diego Sal +4Privacy-Preserving Federated Learning Quantum Federated Learning
Q-AGNN: Quantum-Enhanced Attentive Graph Neural Network for Intrusion Detection
Mar 23, 2026Devashish Chaudhary, Sutharshan Rajasegarar, Shiva Raj PokhrelNetwork Intrusion Detection Graph Attention Networks
School of Information Technology, Deakin University, Geelong, Victoria, 3220, Australia.
Comparative Study of Quantum and Classical Machine Learning Models in Binary Classification
Sep 20, 2026Anand Kumar Mishra, Ramanuj AwasthiQuantum Machine Learning Quantum Kernel Methods
School of Engineering & Technology (UIET), CSJM University, Kanpur, Uttar Pradesh, India

Sep 18, 2026J/K move · Enter open · S save

Federating Quantum and Classical Computing: A Privacy-Preserving Hybrid Approach

Carlos Cano, Daniel M. Jimenez-Gutierrez, Diego Sal, Georgios Kellaris+3

Privacy-Preserving Federated Learning Quantum Federated Learning Quantum Machine Learning Hybrid Quantum-Classical ML Federated Learning

Open paper AskPDF arXiv

Quantum machine learning (QML) is increasingly recognized as one of the most promising near-term applications of quantum computing, viewed as a next-frontier candidate beyond purely classical approaches. Hybrid quantum-classical models operationalize this potential by embedding a parameterized quantum circuit within a model where all other components remain classical-a design already applied to chemistry simulation, financial modeling, and image classification. However, their deployment in privacy-sensitive, multi-party settings is constrained by the need to avoid centralizing raw data and by the requirement that modern quantum circuits remain parameter-efficient to stay trainable at scale. In this paper, we address these constraints by evaluating federated learning (FL) as a means of combining a hybrid quantum-classical active party with a classical passive party, using Sherpa.ai's Blind Vertical FL (SBVFL) protocol to avoid centralizing raw data, while drastically reducing communication. We construct the split multiplicative periodic parity (SMPP) benchmark, following common QML design practice. On this task, our simulations show that SBVFL raises accuracy from 0.7227 to 0.8757 compared to local training, closely approaching non-private centralized accuracy, and that the hybrid quantum-classical model achieves this with substantially fewer trainable parameters than the classical neural networks and random forest alternatives. These results show that FL enables high-performing, privacy-preserving quantum-classical collaboration without centralizing raw data.

Abstract

Quantum Machine Learning

Explore similar work

Federating Quantum and Classical Computing: A Privacy-Preserving Hybrid Approach

Q-AGNN: Quantum-Enhanced Attentive Graph Neural Network for Intrusion Detection

Comparative Study of Quantum and Classical Machine Learning Models in Binary Classification