{"id":"arxiv/2604.24597","url":"https://arxiv.org/abs/2604.24597","tags":["quant-ph","cs.AI"],"score":0,"title":"Quantum Kernel Advantage over Classical Collapse in Medical Foundation Model Embeddings","domain":"quant-ph","source":"arxiv","authors":["Sebastian Cajas Ordóñez","Felipe Ocampo Osorio","Dax Enshan Koh","Rafi Al Attrach","Aldo Marzullo","Ariel Guerra-Adames","J. Alejandro Andrade","Siong Thye Goh","Chi-Yu Chen","Rahul Gorijavolu","Xue Yang","Noah Dane Hebdon","Leo Anthony Celi"],"reviews":0,"subject":"quant-ph","summary":"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…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-04-27","reactions":{}}