{"doi":"10.1109/qce65121.2025.00221","title":"Stochastic Entanglement Configuration for Constructive Entanglement Topologies in Quantum Machine Learning with Application to Cardiac MRI","abstract":"Efficient entanglement strategies are critical for advancing variational quantum circuits (VQCs) in hybrid quantum-classical neural networks (QNNs). However, existing approaches predominantly rely on fixed entanglement topologies that are not adaptive to task-specific requirements, limiting their potential to outperform classical models. To address this limitation, we propose a stochastic entanglement configuration method that systematically and adaptively generates diverse entanglement topologies to identify a subspace of constructive entanglement configurations-defined here as configurations that enhance hybrid model performance (e.g., classification accuracy) relative to classical baselines. In our method, each entanglement configuration is represented as a stochastic binary matrix, where entries denote the presence of directed entanglement between qubits. This formulation enables scalable exploration of the entanglement design space using two key metrics: entanglement density (the proportion of entangled qubits) and per-qubit constraints (the number of entanglements initiated per qubit). We define two sampling modes: an unconstrained mode, allowing variable entanglement per qubit, and a constrained mode, which enforces fixed entanglement per qubit. Using the proposed method, we generated 400 stochastic entanglement configurations and applied them to a hybrid QNN for cardiac magnetic resonance imaging (MRI)-based disease classification. Our technique successfully identified 64 (16%) constructive configurations that consistently outperformed the classical baseline. Ensemble aggregation of top-performing configurations achieved a classification accuracy of <tex xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">$\\sim 0.92$</tex>, outperforming the classical model <tex xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">$(\\sim 0.87)$</tex> by over 5%. Remarkably, when compared to four conventional entanglement topologies/configurations (ring, nearest-neighbor, no entanglement, and fully entangled), none of these configurations surpassed classical performance (maximum accuracy <tex xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">$\\sim 0.82$</tex>), while our identified configurations delivered up to <tex xmlns:mml=\"http://www.w3.org/1998/Math/MathML\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">$\\sim 20 \\%$</tex> higher accuracy performance-highlighting the robustness and generalizability of the newly identified constructive entanglement configurations.","journal":null,"year":2025,"id":583774,"datarank":0.10397207708399181,"base_score":0.6931471805599453,"endowment":0.6931471805599453,"self_citation_contribution":0.10397207708399181,"citation_network_contribution":0.0,"self_endowment_contribution":0.10397207708399181,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":1,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":0.9545,"is_data_producer":false,"deposit_databanks":null,"is_oa":false,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":"2025-01-01","fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":493629,"name":"Mohammed S.M. Elbaz","orcid":"0000-0001-5051-5668","position":1,"is_corresponding":false},{"id":1237484,"name":"Mehri Mehrnia","orcid":"0009-0009-9106-5072","position":0,"is_corresponding":true}],"reference_count":29,"raw_metadata":{"citation_network_status":"fetched"},"created_at":"2026-07-19T02:59:07.970345Z","pmid":null,"pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}