Search papers, labs, and topics across Lattice.
MIT CSAILAI for ResponsibleBeth Israel Deaconess Medical CenterBordeaux Population Health Research CenterClinical Research CenterINRIAJHUNTU TaiwanPoliMiResearch Center of Intelligent Information Processing and Quantum Intelligent ComputingSchool of Information EngineeringSMUUniversidad del CaucaApr 27, 2026arXiv:2604.24597This paper investigates whether quantum support vector machines (QSVMs) can outperform classical linear SVMs in binary insurance classification using embeddings from medical foundation models applied to MIMIC-CXR chest radiographs. They propose a two-tier fair comparison framework where both classifiers receive identical PCA-transformed features. The key finding is that QSVMs consistently outperform classical linear SVMs, which exhibit a "classical collapse" to majority-class prediction, demonstrating a quantum kernel advantage even without hyperparameter tuning.
Quantum kernels unlock signal in medical image embeddings where classical methods fail, suggesting a new path for extracting value from medical foundation models.
We provide evidence of quantum kernel advantage under noiseless simulation in binary insurance classification on MIMIC-CXR chest radiographs using quantum support vector machines (QSVM) with frozen embeddings from three medical foundation models (MedSigLIP-448, RAD-DINO, ViT-patch32). We propose a two-tier fair comparison framework in which both classifiers receive identical PCA-q features. At Tier 1 (untuned QSVM vs. untuned linear SVM, C = 1 both sides), QSVM wins minority-class F1 in all 18 tested configurations (17 at p<0.001, 1 at p<0.01). The classical linear kernel collapses to majority-class prediction on 90-100% of seeds at every qubit count, while QSVM maintains non-trivial recall. At q = 11 (MedSigLIP-448 plateau center), QSVM achieves mean F1 = 0.343 vs. classical F1 = 0.050 (F1 gain = +0.293, p<0.001) without hyperparameter tuning. Under Tier 2 (untuned QSVM vs. C-tuned RBF SVM), QSVM wins all seven tested configurations (mean gain +0.068, max +0.112). Eigenspectrum analysis reveals quantum kernel effective rank reaches 69.80 at q = 11, far exceeding linear kernel rank, while classical collapse remains C-invariant. A full qubit sweep reveals architecture-dependent concentration onset across models. Code: https://github.com/sebasmos/qml-medimage
