Automated Machine Learning can Classify Bound Entangled States with Tomograms

Abstract

For quantum systems with a total dimension greater than six, the positive partial transposition (PPT) criterion is sufficient but not necessary to decide the non-separability of quantum states. Here, we present an Automated Machine Learning approach to classify random states of two qutrits as separable or entangled using enough data to perform a quantum state tomography, without any direct measurement of its entanglement. We could successfully apply our framework even when the Peres-Horodecki criterion fails. In addition, we could also estimate the Generalized Robustness of Entanglement with regression techniques and use it to validate our classifiers.