Figure 3.
Schematic of NN-based QST. (a) Obtain the measured frequencies
and the ground truth of
via Cholesky decomposition for state
. (b) A multilayer NN model to map the frequency vector
to a predicted real vector
, and the NN model is trained to minimize the mean squared error between the predicted
and the expected value
. (c) Obtain a physical density matrix
from the optimal
. The notation Arg indicates the complex phase (in radians) of each density matrix element.
