摘要
We present a machine learning-based approach for efficient quantum state tomography, reconstructing covariance matrices of single- and two-mode squeezed states from homodyne data. This method ensures precise, real-time quantum property estimation and noise robustness.