{"id":"arxiv/2607.11095","url":"https://arxiv.org/abs/2607.11095","tags":["quant-ph","cs.CR","cs.LG"],"score":0,"title":"When cheap gradients fail: the measurement cost of attacking quantum classifiers","domain":"quant-ph","source":"arxiv","authors":["Bacui Li","Chandra Thapa","Tansu Alpcan","Udaya Parampalli"],"reviews":0,"subject":"quant-ph","summary":"Adversarial perturbations threaten machine learning classifiers, including variational quantum classifiers. We show that finite quantum measurement statistics (shot noise) act as a built-in defense against gradient-based test-time attacks whose cost scales unfavorably for the attacker. Because every gradient component must be inferred from repeated circuit executions under any unbiased gradient-estimation rule, white-box extraction consumes a dimension-dependent measurement budget that measurement grouping cannot remove in expressive circuits. Under stated assumptions, single-step attacks…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-07-13","reactions":{}}