Fig. 2: The entanglement estimated by machine learning.
From: Direct entanglement detection of quantum systems using machine learning
a The correlation figures between the predicted entanglement \({{\mathcal{E}}}_{{\rm{ML}}}\) from 2-local measurements and the theoretical \({\mathcal{E}}\). The inset figures are the distributions of the difference \({{\mathcal{E}}}_{{\rm{ML}}}-{\mathcal{E}}\). b Prediction of entanglement dynamics \({{\mathcal{E}}}_{{\rm{ML}}}(t)\) from single-qubit time traces. In each pair of comparisons, the right column is the prediction result obtained by our machine learning method and the left column is the theoretical values. The input layer contains only the measured single-qubit time traces in [0, π]. The trained model allows us to predict \({{\mathcal{E}}}_{{\rm{ML}}}(t)\) at the unseen time [π, 2π]. The measured subsystems are represented by the gray rounded schematic.
