You Only Measure Once: On Designing Single-Shot Quantum Machine Learning Models

Quantum machine learning (QML) models conventionally rely on repeated measurements (shots) of observables to obtain reliable predictions. This dependence on large shot budgets leads to high inference cost and time overhead, which is particularly problematic as quantum hardware access is typically priced proportionally to the number of shots. In this work we propose Y ou Only Measure Once (Y omo), a simple yet effective design that achieves accurate inference with dramatically fewer measurements, down to the single-shot regime. Y omo replaces Pauli expectation-value outputs with a probability aggregation mechanism and introduces loss functions that encourage sharp predictions. Our theoretical analysis shows that Y omo avoids the shot-scaling limitations inherent to expectation-based models, and our experiments on MNIST and CIFAR-10 confirm that Y omo consistently outperforms baselines across different shot budgets and under simulations with depolarizing channels. By enabling accurate single-shot inference, Y omo substantially reduces the financial and computational costs of deploying QML, thereby lowering the barrier to practical adoption of QML. Quantum computing (Nielsen & Chuang, 2010) has emerged as a promising paradigm for advancing computational capabilities beyond the classical regime. Unlike classical machine learning, however, QML inherently involves probabilistic measurement outcomes. To obtain reliable outputs, QML models typically require repeated circuit executions, aggregating many measurement shots to estimate expectation values of observables. This reliance on repeated measurements constitutes one of the fundamental distinctions between classical and quantum machine learning.