Paper links spectral queries to inductive bias in quantum circuits
Original title: Task-aligned spectral inductive biases in parameterized quantum circuits
Encoding a fixed spectral query changes how existing signal frequencies combine without adding sample-dependent information.
The paper studies data-reuploading parameterized quantum circuits that take a varying signal and a fixed, task-defined spectral query. Encoding the query changes the hypothesis class over the signal by modulating existing frequencies rather than introducing new sample-dependent information. On small spectral and synthetic-aperture-radar-inspired tasks, aligned queries improved learning from limited data versus parameter-matched circuits with mismatched or absent queries.
- The mechanism matches tasks based on nonuniform Fourier transforms.
- A radar-like task classifies unfocused spectral patches before image formation.
- Aligned queries beat mismatched or absent queries on small-scale tasks.
- Results are limited to small-scale spectral and radar-like experiments.
From the original article
1 min read at arxiv.orgParameterized quantum circuits are widely investigated in quantum machine learning, yet how their architecture can provide task-aligned inductive biases remains poorly understood. We study data reuploading quantum circuits with two distinct inputs: a signal that varies between samples and a fixed, task-defined spectral query that specifies how the signal should be probed.
By Felix Frohnert · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.