Mixed Quantum States Datasets for Entanglement Classification

Dataset for Supervised Machine Learning Classification of Entanglement in Mixed Quantum States This dataset is designed for the training and evaluation of supervised machine learning algorithms aimed at classifying quantum states. Specifically, the dataset contains labeled examples of 2-qubit and 3-qubit mixed quantum states, categorized as factorizable, separable, and states with negative partial transpose. The data is provided in `.npy` format, compatible with common scientific computing libraries such as NumPy. Each sample represents a quantum state and is labeled according to its classification, enabling the development and benchmarking of models for quantum state discrimination. Data Access and Structure Upon loading, the dataset is structured as a dictionary with two main keys: - 'matrices': contains the quantum state matrices, stored as NumPy arrays of dtype complex128;- 'labels': contains the corresponding classification labels. Labeling Scheme For general classification tasks, classes are ordered as: Fact (factorizable): 0 Sep (separable): 1 NPT (negative partial transpose for all possible bipartitions): 2 For fine-grained labeling of separable classes (e.g., ['NPT', 'Fact', 'Sep_1', 'Sep_2', 'Sep_3']), classes are ordered as: 'Fact': 0 'Sep_1': 1 'Sep_2': 2 'Sep_3': 3 'NPT': 4