Fourier Symmetrization for Geometric Quantum Machine Learning
Abstract
Geometric quantum machine learning incorporates symmetry into quantum models, but how symmetry shapes their expressivity and guides effective model design remains insufficiently understood. We address this question through the Fourier representation of quantum Fourier models (QFMs). Symmetry organizes the frequency spectrum into orbits and sums the Fourier coefficients within each orbit into a symmetrized coefficient. When QFM trainable layers form independent exact 2-designs, the variance of each symmetrized coefficient equals the sum of the coefficient variances in its orbit. For single-layer QFMs with ε-approximate 2-design trainable layers, we bound the deviation from this identity. The resulting bound for individual Fourier coefficients can be exponentially tighter than an existing bound. The hyperoctahedral group provides an example of orbit growth that can mitigate vanishing expressivity of the symmetrized coefficients. Symmetrization of QFMs over an elementary abelian 2-group also yields pure multivariate Chebyshev polynomial basis functions. We introduce randomized encoding, which implements invariant models without ancilla qubits or the additional circuit depth for quantum twirling. We evaluate the models as quantum physicsinformed neural networks (QPINNs) on two-dimensional screened Poisson and stationary viscous Hamilton-Jacobi equations. Under hard boundary constraints, QPINNs using exact symmetrization and randomized encoding achieve the lowest mean errors in the two benchmarks, respectively.
Related
- Fourier-Geometric Circuit Design for Gate and Entanglement Placement in Quantum Neural Networks
- Exoskeleton-based user's pose estimation: a hybrid geometric and machine learning approach
- When do machine-learning potentials need a configuration-dependent kernel? Learning long-range interactions with Fourier neural operators
- Nearly Optimal Quantum Sample Complexity of Learning Fourier-Sparse Boolean Functions
- Hierarchical Fourier Approximation for Variational Quantum Distribution Learning
BibTeX
@misc{wang2026fourier,
title = {{Fourier Symmetrization for Geometric Quantum Machine Learning}},
author = {Letao Wang and Abdel Lisser and Sreejith Sreekumar and Zeno Toffano},
howpublished = {HAL},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.01874},
eprint = {2610.01874},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05774861v1},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Fourier Symmetrization for Geometric Quantum Machine Learning
AU - Letao Wang
AU - Abdel Lisser
AU - Sreejith Sreekumar
AU - Zeno Toffano
PY - 2026
DA - 2026/10/01
PB - HAL
DO - 10.48550/arxiv.2610.01874
UR - https://hal.science/hal-05774861v1
AB - Geometric quantum machine learning incorporates symmetry into quantum models, but how symmetry shapes their expressivity and guides effective model design remains insufficiently understood. We address this question through the Fourier representation of quantum Fourier models (QFMs). Symmetry organizes the frequency spectrum into orbits and sums the Fourier coefficients within each orbit into a symmetrized coefficient. When QFM trainable layers form independent exact 2-designs, the variance of each symmetrized coefficient equals the sum of the coefficient variances in its orbit. For single-layer QFMs with ε-approximate 2-design trainable layers, we bound the deviation from this identity. The resulting bound for individual Fourier coefficients can be exponentially tighter than an existing bound. The hyperoctahedral group provides an example of orbit growth that can mitigate vanishing expressivity of the symmetrized coefficients. Symmetrization of QFMs over an elementary abelian 2-group also yields pure multivariate Chebyshev polynomial basis functions. We introduce randomized encoding, which implements invariant models without ancilla qubits or the additional circuit depth for quantum twirling. We evaluate the models as quantum physicsinformed neural networks (QPINNs) on two-dimensional screened Poisson and stationary viscous Hamilton-Jacobi equations. Under hard boundary constraints, QPINNs using exact symmetrization and randomized encoding achieve the lowest mean errors in the two benchmarks, respectively.
ER -