AI 中文总结
该研究提出量子执行局部性框架（QELF），识别混合量子-经典工作流的五种计算模式，为混合量子-经典计算的推理及相关架构协同设计奠定基础。
AI 中文摘要
混合量子-经典计算已成为近期量子应用的主导范式，但混合工作流通常仅通过单个算法描述，而非其底层执行行为。我们提出量子执行局部性框架（Quantum Execution Locality Framework，QELF），这是一种定性框架，用于根据反复出现的数据流结构和量子执行局部性（即计算在主机干预或经典同步前保持在量子处理单元（Quantum Processing Unit，QPU）上的程度）来表征混合量子-经典工作流。从代表性的应用横截面来看，QELF 识别出五种具有不同局部性特征的反复出现的计算模式，并讨论了它们对通信开销、工作流组织以及未来混合计算架构的影响。通过为混合工作负载的推理提供通用词汇，QELF 为未来的定量验证以及算法、运行时系统和混合计算架构的协同设计奠定了基础。
英文摘要
Hybrid quantum-classical computing has emerged as the dominant paradigm for near-term quantum applications, yet hybrid workflows are typically described by individual algorithms rather than their underlying execution behavior. We introduce the Quantum Execution Locality Framework (QELF), a qualitative framework for characterizing hybrid quantum-classical workflows according to recurring dataflow structures and quantum execution locality, the extent to which computation remains resident on the Quantum Processing Unit (QPU) before host intervention or classical synchronization. From a representative cross-section of applications, QELF identifies five recurring computational patterns with distinct locality characteristics and discusses their implications for communication overhead, workflow organization, and future hybrid computing architectures. By providing a common vocabulary for reasoning about hybrid workloads, QELF establishes a foundation for future quantitative validation and the co-design of algorithms, runtime systems, and hybrid computing architectures.
Comments4 pages. Accepted to the IEEE Quantum Week 2026 Workshop on Advancing Hybrid Computing Through Shared Challenges (WKS115)
