We present the application of KeplerQ-Max, a hybrid quantum–classical machine learning platform built on the Quantum Approximate Optimization Algorithm (QAOA), to the problem of underwater acoustic target recognition using passive sonar data. The system employs a 25-qubit QAOA ansatz executed on IBM Heron R2 real quantum hardware (ibm_fez) to learn discriminative feature representations across three threat classes: Hostile, Unknown, and Friendly. Applied to a dataset of 5,000 sonar records, KeplerQ-Max completed full inference in 64 seconds (12.8 ms per record), classifying 73.58% of targets as Friendly, 6.14% as Hostile, and 20.28% as Unknown, with the Unknown class achieving a top posterior confidence of 98.9%. A total of 1,321 records (26.42%) were flagged as requiring operational action. The Hostile class achieved a mean quantum circuit score of 0.684 and a peak score of 0.806, while 101 high-confidence Unknown records (confidence ≥ 0.90) were identified as priority candidates for secondary sensor fusion. The model (kepler-qmax/qaoa-pytorch-25q9l) integrates a quantum-parameterized variational circuit with a PyTorch post-processing layer, enabling noise-resilient inference on near-term quantum devices at approximately 78 records per second. These results demonstrate that QAOA-based quantum machine learning achieves practical classification throughput on real sonar data without classical simulation, opening pathways for quantum-enhanced naval threat assessment and maritime security applications. Email: research@sonicium.ltd
We present the application of KeplerQ-Max, a hybrid quantum–classical machine learning platform built on the Quantum Approximate Optimization Algorithm (QAOA), to the problem of underwater acoustic target recognition using passive sonar data. The system employs a 25-qubit QAOA ansatz executed on IBM Heron R2 real quantum hardware (ibm_fez) to learn discriminative feature representations across three threat classes: Hostile, Unknown, and Friendly. Applied to a dataset of 5,000 sonar records, KeplerQ-Max completed full inference in 64 seconds (12.8 ms per record), classifying 73.58% of targets as Friendly, 6.14% as Hostile, and 20.28% as Unknown, with the Unknown class achieving a top posterior confidence of 98.9%. A total of 1,321 records (26.42%) were flagged as requiring operational action. The Hostile class achieved a mean quantum circuit score of 0.684 and a peak score of 0.806, while 101 high-confidence Unknown records (confidence ≥ 0.90) were identified as priority candidates for secondary sensor fusion. The model (kepler-qmax/qaoa-pytorch-25q9l) integrates a quantum-parameterized variational circuit with a PyTorch post-processing layer, enabling noise-resilient inference on near-term quantum devices at approximately 78 records per second. These results demonstrate that QAOA-based quantum machine learning achieves practical classification throughput on real sonar data without classical simulation, opening pathways for quantum-enhanced naval threat assessment and maritime security applications. Email: research@sonicium.ltd
The convergence between artificial intelligence (AI) and quantum computing is ushering in a new era for problems requiring high levels of computational complexity, particularly in digital signal processing. This paper presents an overview and evaluation of the development trends in hybrid Quantum-AI models, focusing on applications for the recognition and classification of underwater acoustic signals—a challenging domain due to instability, high noise levels, and data scarcity. The study proposes a Hybrid Quantum-CNN model, utilizing CNN for feature extraction and dimensionality reduction, combined with a Variational Quantum Classifier for classification in Hilbert space. The results of the proposed model are compared with an equivalent CNN network, demonstrating that HQC achieves higher accuracy (an increase of 3,5%), while excelling in efficiency: faster convergence speed, and superior generalization capabilities on the same real underwater acoustic dataset. These findings highlight the benefits of quantum approaches not only in accuracy but also in constructing more efficient and robust models; hybrid models open promising avenues for applying quantum machine learning to complex signal reception systems in practical.
In the big data era, traditional machine learning algorithms like K-means face computational challenges in processing large, high-dimensional datasets, particularly for underwater acoustic signal classification. This study investigates the application of the quantum clustering algorithm Q-means, a quantum-enhanced variant of K-means, to real passive sonar datasets. The purpose is to evaluate Q-means' effectiveness in classifying propeller-driven ship signals under noisy conditions, leveraging quantum principles such as superposition and entanglement for exponential speedup. The research extends classical K-means to δ-k-means for robustness, implementing quantum subroutines including distance estimation, cluster assignment, and state tomography. Preprocessed features from power spectral density analysis of real datasets (from prior studies on varying ship speeds) are used as input to avoid issues with raw time-series data, such as high dimensionality and sensitivity to shifts. Simulations on 15,000 passive sonar samples demonstrate that Q-means achieves clustering quality comparable to K-means, with clear separation of three clusters and accurate centroids, while reducing complexity from O(n) to O(polylog(n)). This validates Q-means as a promising tool for large-scale, noisy acoustic data in national security applications, bridging quantum theory with practical machine learning.
This study investigates classification performance and trainability of variational quantum classifiers on near-term noisy quantum hardware across circuit depth and qubit count regimes within the context of quantum computing and machine learning, an area of growing scientific importance given its implications for hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment. Using gradient-based parameter optimization of parameterized quantum circuits with hardware noise modeling via density matrix simulation and physical execution on IBM Eagle processors, we examine expressibility of parameterized circuits enabling nonlinear decision boundaries in feature Hilbert space, with trainability constrained by barren plateau gradients at large qubit counts in 96 circuit architectures (4 qubit counts x 4 depths x 6 ansatz families) evaluated across 5 classification datasets with 20 random initialization seeds each drawn from IBM Quantum Eagle (127-qubit) and Falcon (27-qubit) processors with noise mitigation via zero-noise extrapolation. Results indicate that 10-qubit, 4-layer VQC achieves 84.2% classification accuracy on the best benchmark, within 3.8 percentage points of classical SVM, but gradient variance decays exponentially with qubit count (variance halving per 2 added qubits), confirming barren plateau scaling (p < 0.001), with 84.2% accuracy, 3.8 pp gap to classical SVM baseline as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to quantum computing and machine learning and carry actionable implications for the design of programs and policies targeting hybrid quantum-classical algorithm design and practical NISQ hardware benchmarking for near-term quantum advantage assessment.
Quantum sonar denoising leverages variational quantum algorithms to suppress noise in underwater acoustic signals beyond classical limits. This study presents the first successful execution of a three-qubit Variational Quantum Eigensolver (VQE) for quantum sonar denoising on real IBM superconducting quantum hardware (ibm_torino, 127-qubit Eagle-class processor). Using a hardware-efficient ansatz with θ = π, the exact ground state of the target Hamiltonian H = 0.5 ZIZ + 0.3 IZI + 0.2 ZII was prepared with perfect analytical energy of –0.800000. Two independent runs were performed: (1) state preparation without measurement (exact energy achieved), and (2) full measurement-enabled run with 16,000 shots yielding real-hardware energy of 0.008088 due to gate and readout noise. Histogram analysis confirms dominant population in the expected ground-state manifold (|000⟩, |111⟩). Results demonstrate 100% theoretical fidelity in state preparation and real-world quantum execution, establishing a milestone in quantum-enhanced signal processing.
Abstract The Quantum approximate optimization algorithm (QAOA) constitutes one of the often mentioned candidates expected to yield a quantum boost in the era of near-term quantum computing. In practice, quantum optimization will have to compete with cheaper classical heuristic methods, which have the advantage of decades of empirical domain-specific enhancements. Consequently, to achieve optimal performance we will face the issue of algorithm selection, well-studied in practical computing. Here we introduce this problem to the quantum optimization domain. Specifically, we study the problem of detecting those problem instances of where QAOA is most likely to yield an advantage over a conventional algorithm. As our case study, we compare QAOA against the well-understood approximation algorithm of Goemans and Williamson on the Max-Cut problem. As exactly predicting the performance of algorithms can be intractable, we utilize machine learning (ML) to identify when to resort to the quantum algorithm. We achieve cross-validated accuracy well over 96%, which would yield a substantial practical advantage. In the process, we highlight a number of features of instances rendering them better suited for QAOA. While we work with simulated idealised algorithms, the flexibility of ML methods we employed provides confidence that our methods will be equally applicable to broader classes of classical heuristics, and to QAOA running on real-world noisy devices.
Abstract Quantum machine learning for genomic sequence classification faces three compounding constraints on current hardware: barren plateaus in variational training, exponential concentration of quantum kernel expressivity in highdimensional feature spaces, and noise accumulation that limits practical circuit depth to fewer than 50 layers on current NISQ devices. Existing quantum genomics studies address these constraints using fixed-architecture circuits without noise-calibrated validation on standardized benchmarks, leaving the potential of adaptive circuit topology largely unexplored. Three QAOA variants are developed and systematically compared to address this gap. Classical QAOA establishes a baseline using fixed p-layer circuits with standard phase and mixing operators. Alternating QAOA introduces structured CNOT entanglement between alternating qubit pairs to evaluate whether non-local correlations improve genomic pattern recognition. The primary contribution, Adaptive QAOA, employs gradient-driven operator selection from feature-weighted Pauli operator pools, generating variable-depth problemspecific circuits that adapt their topology to dataset structure rather than applying universal ans¨atze. All three architectures are evaluated through quantum simulation on two standardized genomic benchmarks: Human Nontata Promoters (36,131 sequences) and Demo Coding vs. Intergenic Sequences (100,000 sequences). Experiments are conducted under both noiseless conditions and noise models calibrated to IBM superconducting hardware specifications, across feature space dimensionalities ranging from 2 to 10 dimensions using 10-fold cross-validation. Adaptive QAOA consistently achieves the highest precision (approximately 75% under noiseless simulation) and maintains perfect specificity across both datasets and noise conditions, results not observed for fixed-depth variants. These performance gains are accompanied by reduced hardware resource consumption, with Adaptive QAOA requiring the fewest qubits and lowest quantum usage time (∼0.52–0.60 s) across most configurations. Superior coherence characteristics — average T1 of 82.12 μs and T2 of 55.45 μs, exceeding Classical QAOA by 13.6% and 113.3% respectively — provide a physical basis for reduced error accumulation under realistic noise constraints. All QAOA variants maintain stable classification performance across all tested dimensionalities, contrasting with compounding errors documented for fixed-architecture quantum kernel methods on high-dimensional genomic data. These results establish that gradient-driven adaptive circuit topology with domain-specific feature weighting represents a more effective design strategy than uniform depth scaling or entanglement-based ansatz modification for near-term quantum genomic classification.