Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a quantum data loading method that learns a shared, low-dimensional structure from real-world datasets, enabling a fixed quantum circuit to prepare any signal within that learned manifold. This approach significantly reduces the number of parameters required per signal compared to existing structured loaders while achieving comparable accuracy on benchmark datasets. The learned structure's efficiency is demonstrated by its ability to infer accurate state preparation from small data subsets, even as signal complexity increases.
A single, learned quantum circuit can prepare diverse real-world signals with orders of magnitude fewer parameters than previously possible.
Preparing quantum states from classical data can cost more than the computation they serve; most loaders tailor a circuit to each input. Here we show that the signals of a real dataset share structure that can be learned once and reused. Our quantum-native loader learns a low-dimensional description of a dataset and prepares every signal with one fixed circuit set by a few numbers. Across seven views of five public datasets it meets the targets of the strongest structured loader at equal gate cost with several times fewer numbers per signal. These numbers can be inferred from a random subset: in a preregistered blind replication the subset needed to come within ten per cent of full-signal accuracy stayed constant within a prespecified margin as signals grew sixteenfold, whereas the structured loader needed ever more. It declines what it cannot represent, covering fewer cases than that baseline and no electrocardiogram.