Neural networks-assisted state tomography for NV-center spin registers

The goal of the thesis is to use Neural Networks to reconstruct spin states in NV-center platforms from the outcomes of computational basis measurements. The use of Artificial Intelligence to study quantum systems is a promising technique that is progressively gaining popularity, especially within the framework of Quantum State Tomography. Neural Networks, in fact, can be seen as a more efficient alternative to standard reconstruction protocols, thanks to their ability to find patterns between data with fewer information. Specifically, we apply these tools to perform tomography of spin states in NV-center registers. This quantum platform consists of an electronic spin of a Nitrogen-Vacancy defect in diamond, used to perform initialization and readout, and the surrounding nuclear Carbon spins, which act as the actual qubits. In the first part of the results, reconstruction is performed from outcomes of Pauli measurements, obtained through a series of gates tailored to encode the information of Pauli operators on a single qubit applied before the computational basis readout. Then, we provide a comparison with a randomized protocol, where randomly selected gates from a closed pool are applied before readout.