Decay versus dephasing in Rydberg analog optimization: the exchange rate depends on the readout
Abstract
Numerical studies of noisy Rydberg-atom optimization compress decoherence into a single rate, treating spontaneous decay (Γ) and dephasing (γ) as equivalent. We resolve the two as independent Lindblad axes for quantum annealing of unit-disk maximum independent set at fixed annealing time and ask how their relative effect depends on the readout. For the raw readout the size error collapses onto the effective dephasing rate u=κΓ+γ, and at small noise a unit of decay rate degrades the result as much as κ=7.4 units of dephasing rate; a collapse fit across the whole grid returns 8.3. After greedy completion followed by a (1,2)-swap local search, classical post-processing of polynomial cost, the exchange rate falls to κ=1.6, and at the calibrated operating point of a commercial device dephasing produces three quarters of the noise-induced error instead of two fifths. First-order perturbation theory built from noiseless propagation gives the decay and dephasing susceptibilities of every readout, raw or post-processed, to a few parts in 106 of the full Lindblad result. The raw asymmetry is set at the end of the schedule, where a decay event deletes an atom and a dephasing event leaves the readout unchanged. Post-processing refills most deletions and leaves misplaced atoms, which the two channels produce at comparable rates. Which channel limits a device therefore depends on the readout as well as on the hardware.
Related
- An exchange-assisted entangling gate between 87Rb and 171Yb Rydberg atoms
- Rydberg-Atom-Mediated Strong Antisymmetric Spin Exchange in Molecular Arrays
- Spin-exchange interactions between circular Rydberg atoms over long times
- Error-Corrected Quantum Optimization via Rydberg Atom Array Embeddings
- Application and Optimization of the GET Electronics System for Double-Sided Silicon Strip Detector Readout
BibTeX
@misc{kim2026decay,
title = {{Decay versus dephasing in Rydberg analog optimization: the exchange rate depends on the readout}},
author = {Seunghyeon Kim and Minju Kim and Junwoo Jung},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2608.29858},
eprint = {2608.29858},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.29858},
}
RIS
TY - UNPB
TI - Decay versus dephasing in Rydberg analog optimization: the exchange rate depends on the readout
AU - Seunghyeon Kim
AU - Minju Kim
AU - Junwoo Jung
PY - 2026
DA - 2026/10/29
PB - arXiv
DO - 10.48550/arxiv.2608.29858
UR - https://arxiv.org/abs/2608.29858
AB - Numerical studies of noisy Rydberg-atom optimization compress decoherence into a single rate, treating spontaneous decay (Γ) and dephasing (γ) as equivalent. We resolve the two as independent Lindblad axes for quantum annealing of unit-disk maximum independent set at fixed annealing time and ask how their relative effect depends on the readout. For the raw readout the size error collapses onto the effective dephasing rate u=κΓ+γ, and at small noise a unit of decay rate degrades the result as much as κ=7.4 units of dephasing rate; a collapse fit across the whole grid returns 8.3. After greedy completion followed by a (1,2)-swap local search, classical post-processing of polynomial cost, the exchange rate falls to κ=1.6, and at the calibrated operating point of a commercial device dephasing produces three quarters of the noise-induced error instead of two fifths. First-order perturbation theory built from noiseless propagation gives the decay and dephasing susceptibilities of every readout, raw or post-processed, to a few parts in 10^6 of the full Lindblad result. The raw asymmetry is set at the end of the schedule, where a decay event deletes an atom and a dephasing event leaves the readout unchanged. Post-processing refills most deletions and leaves misplaced atoms, which the two channels produce at comparable rates. Which channel limits a device therefore depends on the readout as well as on the hardware.
ER -