Quantum Feature Amplification Network (QFAN) as An Autoregressive Quantum Generative Model
Abstract
Gate-based quantum generative models have been limited to small images, because their quantum resources grow with the size of the data. The Quantum Feature Amplification Network (QFAN) removes this limit. It reuses a single few-qubit circuit to generate the data block by block, each block conditioned on a compact summary of those before it, with the circuit's measurement outcomes as the source of randomness. The number of qubits is set by the block rather than by the data, and the number of circuits per training step does not grow with the data either. In simulations with only four qubits, QFAN generates full-resolution calorimeter showers from the CaloChallenge Dataset~1 benchmark, 368 voxels each, and learns their correlations and shower-to-shower fluctuations. A three-qubit version generates 12- and 25-pixel images on IBM quantum hardware. This brings high-dimensional scientific data within reach of near-term quantum generative models.
Related
- Hybrid Machine Learning-Assisted Raman Spectroscopy with Generative Feature Augmentation for Pharmaceutical Identification
- High-speed optical target feature extractor based on multimode fiber optical neural network
- Optimal Quantum Purity Amplification: A Top-Down Approach to Error Suppression via Symmetry and Reversibility
- Quantum Noise Limited Nonlinear Phase-Preserving Amplification of a Bosonic Mode
- Making the most of leftovers: Improved privacy amplification for quantum key distribution
BibTeX
@misc{slim2026quantum,
title = {{Quantum Feature Amplification Network (QFAN) as An Autoregressive Quantum Generative Model}},
author = {Jamal Slim and Saverio Monaco and Florian Rehm and Dirk Kruecker and Kerstin Borras},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2605.16044},
eprint = {2605.16044},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.16044},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Quantum Feature Amplification Network (QFAN) as An Autoregressive Quantum Generative Model
AU - Jamal Slim
AU - Saverio Monaco
AU - Florian Rehm
AU - Dirk Kruecker
AU - Kerstin Borras
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2605.16044
UR - https://arxiv.org/abs/2605.16044
AB - Gate-based quantum generative models have been limited to small images, because their quantum resources grow with the size of the data. The Quantum Feature Amplification Network (QFAN) removes this limit. It reuses a single few-qubit circuit to generate the data block by block, each block conditioned on a compact summary of those before it, with the circuit's measurement outcomes as the source of randomness. The number of qubits is set by the block rather than by the data, and the number of circuits per training step does not grow with the data either. In simulations with only four qubits, QFAN generates full-resolution calorimeter showers from the CaloChallenge Dataset~1 benchmark, 368 voxels each, and learns their correlations and shower-to-shower fluctuations. A three-qubit version generates 12- and 25-pixel images on IBM quantum hardware. This brings high-dimensional scientific data within reach of near-term quantum generative models.
ER -