Quantum Evolution Networks Learn Long-Time Coherent Dynamics in Driven Spin Registers
Abstract
Predicting and controlling driven quantum systems requires access to dynamics across physical parameters, initial states, and evolution times. A reusable neural model must capture this dependence while preserving unitary structure and maintaining accuracy over long-time coherent evolution. Here we introduce Quantum Evolution Networks (QENs), which map physical parameters and time to complete propagators for periodically driven three-spin registers. Shared networks learn unitary one-period propagators and compatible within-period paths, enabling predictions through known periodic composition without query-specific numerical integration. For transverse driving, the fifth-percentile fidelity, minimized over input states and sampled times, is 0.9947 over 24 periods on 256 held-out configurations. A leading-order analysis of spectrally resolved endpoint residuals predicts multicycle process error more accurately than quadratic or spectrum-averaged estimates. Interventions at matched single-period infidelity produce orders-of-magnitude differences in multicycle infidelity, demonstrating the importance of residual orientation. Compatible component replacements identify how endpoint and within-period approximations limit reconstruction accuracy. These results establish a reusable neural representation of the studied propagator families and connect its learned residual structure to its observed long-time reliability.
Related
- Quantum Hamiltonian Evolution for Coherent Quantum Learning
- Coherent quantum state corrections in RN–black holes: Photon dynamics, shadow casts, and energy backreaction
- Terahertz two-dimensional coherent spectroscopy: unlocking coherent quantum intraband nonlinear dynamics
- Coherent dynamics of individual excitons in a quantum dot embedded in a nanopost
- Modeling microbiome dynamics on social networks: how between-host transmission shapes within-host evolution
BibTeX
@misc{li2026quantum,
title = {{Quantum Evolution Networks Learn Long-Time Coherent Dynamics in Driven Spin Registers}},
author = {Puran Li and Condon Lau and Xiu-Hao Deng},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.32156},
eprint = {2610.32156},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.32156},
}
RIS
TY - UNPB
TI - Quantum Evolution Networks Learn Long-Time Coherent Dynamics in Driven Spin Registers
AU - Puran Li
AU - Condon Lau
AU - Xiu-Hao Deng
PY - 2026
DA - 2026/10/23
PB - arXiv
DO - 10.48550/arxiv.2610.32156
UR - https://arxiv.org/abs/2610.32156
AB - Predicting and controlling driven quantum systems requires access to dynamics across physical parameters, initial states, and evolution times. A reusable neural model must capture this dependence while preserving unitary structure and maintaining accuracy over long-time coherent evolution. Here we introduce Quantum Evolution Networks (QENs), which map physical parameters and time to complete propagators for periodically driven three-spin registers. Shared networks learn unitary one-period propagators and compatible within-period paths, enabling predictions through known periodic composition without query-specific numerical integration. For transverse driving, the fifth-percentile fidelity, minimized over input states and sampled times, is 0.9947 over 24 periods on 256 held-out configurations. A leading-order analysis of spectrally resolved endpoint residuals predicts multicycle process error more accurately than quadratic or spectrum-averaged estimates. Interventions at matched single-period infidelity produce orders-of-magnitude differences in multicycle infidelity, demonstrating the importance of residual orientation. Compatible component replacements identify how endpoint and within-period approximations limit reconstruction accuracy. These results establish a reusable neural representation of the studied propagator families and connect its learned residual structure to its observed long-time reliability.
ER -