REVIEW 1 cited by
SantaQlaus: A resource-efficient method to leverage quantum shot-noise for optimization of variational quantum algorithms
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
SantaQlaus: A resource-efficient method to leverage quantum shot-noise for optimization of variational quantum algorithms
read the original abstract
We introduce SantaQlaus, a resource-efficient optimization algorithm tailored for variational quantum algorithms (VQAs), including applications in the variational quantum eigensolver (VQE) and quantum machine learning (QML). Classical optimization strategies for VQAs are often hindered by the complex landscapes of local minima and saddle points. Although some existing quantum-aware optimizers adaptively adjust the number of measurement shots, their primary focus is on maximizing gain per iteration rather than strategically utilizing quantum shot-noise (QSN) to address these challenges. Inspired by the classical Stochastic AnNealing Thermostats with Adaptive momentum (Santa) algorithm, SantaQlaus explicitly leverages inherent QSN for optimization. The algorithm dynamically adjusts the number of quantum measurement shots in an annealing framework: fewer shots are allocated during the early, high-temperature stages for efficient resource utilization and landscape exploration, while more shots are employed later for enhanced precision. Numerical simulations on VQE and QML demonstrate that SantaQlaus outperforms existing optimizers, particularly in mitigating the risks of converging to poor local optima, all while maintaining shot efficiency. This paves the way for efficient and robust training of quantum variational models.
Forward citations
Cited by 1 Pith paper
-
When cheap gradients fail: the measurement cost of attacking quantum classifiers
Unbiased gradient extraction for attacking quantum classifiers costs at least Θ(d^{5/2}) shots under norm-concentration scaling, and ~d³ for tested deep circuits, so the attacker's relative cost diverges versus classi...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.