Operator-Guided Model Reduction for Generative Sampling in Lattice Field Theory
Abstract
Neural generative samplers for lattice field theory can be costly to train and evaluate. When they miss modes or assign them incorrect relative weights, biased observables do not reveal which collective variables are responsible. We project a trained flow-matching velocity onto vector fields built from lattice operators and Fourier modes. In two-dimensional lattice ϕ4 theory, the projection separates changes in the overall magnetization from the lowest nonzero-momentum fluctuations and guides an explicit invertible proposal that treats them separately. Allowing the amplitude of the lowest nonzero-momentum fluctuations to depend on the magnetization improves the overlap between the proposal and target distributions, while the same two-parameter modification at higher momenta gives smaller improvements. The Metropolis--Hastings correction defines a Markov chain with the target Boltzmann distribution as its stationary law, and the normalized proposal density yields finite-volume partition-function estimates consistent with an independent HMC calculation. At the larger tested volume, the overlap between the proposal and target distributions deteriorates substantially, limiting the range over which the same parameterization remains effective.
Related
- Generative sampling with physics-informed kernels
- Generative learning of time series with holographic quantum circuit sampling
- On Certifying Source Sampling Hardness in Quantum Generative Modeling
- Physics-guided topology-aware safeguarded unfolding for low-sampling computational ghost imaging
- Verifiable Sampling for Delegated Generative Quantum Machine Learning
BibTeX
@misc{qian2026operatorguided,
title = {{Operator-Guided Model Reduction for Generative Sampling in Lattice Field Theory}},
author = {Moxian Qian},
howpublished = {arXiv},
year = {2026},
month = aug,
doi = {10.48550/arxiv.2605.11199},
eprint = {2605.11199},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.11199},
}
RIS
TY - UNPB
TI - Operator-Guided Model Reduction for Generative Sampling in Lattice Field Theory
AU - Moxian Qian
PY - 2026
DA - 2026/08/30
PB - arXiv
DO - 10.48550/arxiv.2605.11199
UR - https://arxiv.org/abs/2605.11199
AB - Neural generative samplers for lattice field theory can be costly to train and evaluate. When they miss modes or assign them incorrect relative weights, biased observables do not reveal which collective variables are responsible. We project a trained flow-matching velocity onto vector fields built from lattice operators and Fourier modes. In two-dimensional lattice ϕ^4 theory, the projection separates changes in the overall magnetization from the lowest nonzero-momentum fluctuations and guides an explicit invertible proposal that treats them separately. Allowing the amplitude of the lowest nonzero-momentum fluctuations to depend on the magnetization improves the overlap between the proposal and target distributions, while the same two-parameter modification at higher momenta gives smaller improvements. The Metropolis--Hastings correction defines a Markov chain with the target Boltzmann distribution as its stationary law, and the normalized proposal density yields finite-volume partition-function estimates consistent with an independent HMC calculation. At the larger tested volume, the overlap between the proposal and target distributions deteriorates substantially, limiting the range over which the same parameterization remains effective.
ER -