This paper studies how spectral geometry emerges in quantum learning models and how it can be diagnosed with physically grounded probes. In graph-regularized quantum networks, training reorganizes the output similarity graph, increases the effective spectral dimension Delta S = +0.23, and reshapes the Laplacian spectrum. Edge-resolved two-boson interference directly probes this restructuring: the bosonic enhancement Delta P_uv correlates with the Fiedler edge split |Delta v_2| (r = -0.50), linking learned spectral partitions to interference signatures. A phase diagram shows a nonmonotonic dependence of performance on coupling strength gamma and noise delta, with graph regularization improving fidelity only in a restricted regime; hardware experiments confirm the predicted interference behavior within shot-noise uncertainty. We also analyze a hybrid quantum autoencoder and introduce Bloch-space drift as a geometric diagnostic of its latent representation. With an unsupervised benign-data threshold, the model achieves high ranking performance (ROC-AUC about 0.99) and negligible false-negative rates. Absolute Bloch drift strongly discriminates anomalies (ROC-AUC at least about 0.9), while consecutive drift is near random (ROC-AUC about 0.5), showing that detection arises from persistent state-space displacement rather than local fluctuations. Through the geometry of reduced single-qubit states and associated quantum Fisher information, these results show that learning-induced spectral organization appears as measurable quantum-state structure, establishing a unified spectral-geometric framework for diagnosing quantum learning systems with bosonic and Bloch probes.
Hilbert-space fragmentation provides a mechanism for ergodicity breaking in quantum many-body systems even in the absence of disorder, leading to dynamically disconnected sectors and strong memory of initial conditions. However, identifying such structures is often challenging and typically relies on prior knowledge of conservation laws or model-specific analytical insight. Here we introduce an unbiased approach based on spectral graph theory and, within this framework, formulate the concept of nearly fragmented systems, in which perturbative processes couple otherwise fragmented sectors while preserving their dynamical imprint. By representing basis states as vertices and Hamiltonian matrix elements as edges, we map the connectivity structure of the many-body Hilbert space onto a graph and analyze it using tools such as the Laplacian spectrum, Fiedler vectors, and modularity. Exact fragmentation corresponds to disconnected graph components, while nearly fragmented systems manifest as weakly connected communities whose structure can still be resolved spectrally. Applying this framework to the one-dimensional
t-
J model and its perturbations, we demonstrate that graph-theoretic diagnostics reliably identify both fragmented and nearly fragmented Hilbert-space structures and capture the hierarchy of dynamical time scales that governs the system's evolution. We further show that the method extends beyond kinetically constrained models by applying it directly to the Hubbard chain, where it reveals the emergence of nearly decoupled subspaces associated with doublon dynamics and spin configurations. Our results establish the spectral graph analysis as a general and scalable tool for diagnosing fragmentation and approximate dynamical constraints in complex quantum many-body systems.
Spectral clustering is a powerful unsupervised machine learning algorithm for clustering data with nonconvex or nested structures [A. Y. Ng, M. I. Jordan, and Y. Weiss, On spectral clustering: Analysis and an algorithm, in Advances in Neural Information Processing Systems 14: Proceedings of the 2001 Conference (MIT Press, Cambridge, MA, 2002), pp. 849--856]. With roots in graph theory, it uses the spectral properties of the Laplacian matrix to project the data in a low-dimensional space where clustering is more efficient. Despite its success in clustering tasks, spectral clustering suffers in practice from a fast-growing running time of
O(n3), where
n is the number of points in the data set. In this work we propose an end-to-end quantum algorithm performing spectral clustering, extending a number of works in quantum machine learning. The quantum algorithm is composed of two parts: the first is the efficient creation of the quantum state corresponding to the projected Laplacian matrix, and the second consists of applying the existing quantum analog of the
k-means algorithm [I. Kerenidis, J. Landman, A. Luongo, and A. Prakash,
q-means: A quantum algorithm for unsupervised machine learning, in Advances in Neural Information Processing Systems 32: Proceedings of the 2019 Conference (Curran Associates, Red Hook, NY, 2020), pp. 4136--4146]. Both steps depend polynomially on the number of clusters, as well as precision and data parameters arising from quantum procedures, and polylogarithmically on the dimension of the input vectors. Our numerical simulations show an asymptotic linear growth with
n when all terms are taken into account, significantly better than the classical cubic growth. This work opens the path to other graph-based quantum machine learning algorithms, as it provides routines for efficient computation and quantum access to the incidence, adjacency, and projected Laplacian matrices of a graph.
The conceptual framework of quantum computing has recently emerged as promising to enhance machine learning in complex non-stationary settings of sensing, but its use is still challenged by hardware noise and concept drift in streaming data. This paper demonstrates a noise-aware quantum kernel method of online anomaly detection on high-velocity sensor streams in changing data distributions. To represent the signal magnitude and dynamics, a two-dimensional temporal embedding is used and quantum kernel evaluations are done via Monte Carlo sampling of explicitly depolarizing and readout noisy models to simulate near-term hardware constraints. The resultant uncertainty is harnessed in confidence-weighted anomaly scoring and adaptive thresholding which allows effective drift-tolerant detection with no labeled data. The suggested approach is tested on streaming sensor benchmarks using injected concept drift and evaluated against radial basis function (RBF) kernels and ideal (noise-free) quantum kernels. Through experimental results, it is shown that although classical RBF has the highest ROC-AUC of 0.9762 and PR-AUC of 0.4444, the proposed quantum noise-aware RBF generates impressive results achieve ROC-AUC of 0.9017 and PR-AUC of 0.4134, which means most of the lost performance is associated with quantum noise. The findings emphasize the significance of the explicit modeling of quantum uncertainty and show the practical feasibility of quantum machine learning in the real-time non-stationary deployment-oriented anomaly detection in sensor systems.
This work investigates whether classical machine learning can detect statistical signatures associated with entanglement graph structure from partial quantum measurements. Experiments were performed on real IBM Quantum hardware using circuits with controlled entanglement patterns. Only a subset of qubits was observed, while the state of one qubit was hidden. A classical Random Forest classifier was trained to predict the hidden qubit’s measurement outcome based solely on the observed subset. To distinguish structural effects from spatial or hardware-related confounders, a geometry decoupling protocol was introduced. This involved comparing entangled circuits with control circuits of identical spatial layout but without entangling gates. Results show that classification accuracy exceeds chance in entangled configurations and correlates negatively with graph distance, suggesting that classical models can capture statistical signatures consistent with entanglement connectivity. The introduced delta metric partially controls for hardware bias by comparing entangled and non-entangled baselines. The observed signal reflects classical correlations in measurement outcomes generated by entangled states and does not constitute quantum communication or violation of no-signalling principles. This dataset and code provide a reproducible experimental framework for studying the relationship between quantum circuit structure and classical detectability using machine learning.
Geometric Machine Learning (GML) successes have been achieved through the thorough study and design of new equivariant neural networks. In comparison, geometric quantum machine learning (GQML) models lack such a detailed understanding and, despite already several proposals, a unifying perspective on their design remains elusive. In this work, we focus on GQML models for graph problems that showcase a lot of structure and still remain frontier in machine learning. For the case when n-node graphs are encoded in n-qubit states, we provide a comprehensive characterization of their constituents. Taken together, these furnish us with a toolbox for the design of quantum graph models, and we further probe its benefits including the natural integration with classical models, generalization of known GQML models (sometimes extending their expressivity at virtually no cost), and straightforward classical pre-training strategies. The latter two features are demonstrated in dedicated numerical experiments.