Fig. 5: Classification of 1D and 2D datasets.
From: Quantum machine learning with Adaptive Boson Sampling via post-selection
a Example of classification of the 1D dataset performed by the SVM with the two quantum experimental kernels obtained from qubit states and qutrit states. The labels yi of the dataset are shown in the colors green and gray, while the symbols ‘o’ and ‘x’ indicate the training and test set, respectively. The background color represents the result of the classification. b–c Histograms of the average accuracy for classification with kernels collected in the experiment by choosing different assignments of the post-selected modes to the adaptive operation. For each kernel, the accuracy is averaged over 100 random partitions of data into training and test sets. In b results for qubits' kernels and in c the qutrits case. d–e Classification of a 2D dataset done with a preliminary clustering algorithm (K-means) followed by the application of a SVM with the two quantum experimental kernels obtained from d qubit states and e qutrit states. The correct label yi of the dataset is shown with the color (green/gray) of the symbols (‘o’: training set, ‘x’: test set). The background color represents the result of the classification.
