用可编程光子学对部分相干光进行与架构无关的分析
Architecture-agnostic analysis of partially coherent light with programmable photonics
浏览论文内容
中文总结 AI 辅助
研究如何精确表征辐射场空间相干度，提出与架构无关的方法，利用舒尔 - 霍恩定理对角化输出相干矩阵，通过输出功率测量提取特征值，经数值验证有效且对损耗和偏差有弹性，为空间相干分析仪发展铺路。
中文摘要 AI 辅助
辐射场空间相干度的精确表征对于评估其在光通信、先进成像和量子信息处理等特定应用中的适用性至关重要。然而，传统上测量完整的相干矩阵需要复杂的、对相位敏感的干涉测量装置，这种装置极易受到噪声影响且难以在集成平台上扩展。为解决此问题，我们提出一种与架构无关的方法来分析部分相干光，该方法与任何通用可编程光子酉电路兼容，无论其内部拓扑结构如何。利用舒尔 - 霍恩定理，我们的方法使输出相干矩阵对角化，仅通过输出功率测量就能直接提取其特征值。我们在各种通用拓扑结构上对该框架进行了数值验证，并证明即使在参数不足的非通用架构中，其精度仅有轻微损失时也有效。最后，我们的黑箱优化方法被证明对任意光学损耗和组件偏差具有内在弹性，为强大、低深度且可编程的空间相干分析仪铺平了道路。
英文摘要
The precise characterization of the spatial degree of coherence of a radiation field is important for assessing its suitability for specific applications in optical communications, advanced imaging, and quantum information processing. However, measuring the full coherence matrix traditionally requires complex, phase-sensitive interferometric setups that are highly susceptible to noise and difficult to scale on integrated platforms. To address this, we propose an architecture-agnostic approach for analyzing partially coherent light that is compatible with any universal programmable photonic unitary circuit, regardless of its internal topology. Leveraging the Schur-Horn theorem, our method diagonalizes the output coherence matrix, enabling direct extraction of its eigenvalues from output power measurements alone. We numerically validate this framework across various universal topologies and demonstrate its efficacy even in under-parameterized, non-universal architectures with only minor loss in precision. Finally, our black-box optimization approach proves inherently resilient to arbitrary optical losses and component deviations, paving the way for robust, lower-depth, and programmable spatial coherence analyzers.
