Queue management of various quantum processing units provided by a quantum computing service

By training machine learning models in quantum computing services to predict the latency of quantum processing units, the problem of unreliable latency in quantum computing services is solved, enabling efficient quantum task management and customized execution.

CN122374768APending Publication Date: 2026-07-10AMAZON TECH INC
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
AMAZON TECH INC
Filing Date
2024-12-05
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Due to the nondeterministic nature of quantum computing and various developmental factors, it is difficult to reliably predict the waiting time of quantum computing resources, resulting in inefficiency in managing and optimizing quantum task queues for quantum computing services.

Method used

Quantum computing services predict the latency of quantum processing units (QPUs) by training machine learning models, utilizing labeled datasets and supervised learning techniques, and combining machine learning models to optimize the selection of quantum task queues and load balancing, providing customized quantum task execution services.

Benefits of technology

It enables efficient management of quantum tasks and accurate prediction of latency, improving the reliability and efficiency of quantum computing services and ensuring the successful execution of customer quantum objects.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122374768A_ABST
    Figure CN122374768A_ABST
Patent Text Reader

Abstract

Techniques for tracking and maintaining queues for executing pending quantum objects using respective quantum processing units (QPUs) are disclosed. The amount of time to execute a given quantum object depends on many factors, and the non-deterministic nature of quantum computing resources makes it difficult to reliably determine, despite knowing that it is useful to access expected wait times in queues for a given QPU. A quantum computing service that manages submission of quantum objects to respective QPUs and execution of quantum objects can apply a QPU-specific machine learning model in order to predict expected wait times and provide that information to customers. The respective machine learning models can be trained using supervised learning techniques by generating a labeled dataset using benchmark real wait times related to quantum objects that have already been executed, which can be a self-contained and recurring process.
Need to check novelty before this filing date? Find Prior Art