In contrast to the full density matrix, the corresponding quantum phase-space distribution of a continuous-variable system can be experimentally accessed. For example that of a quantum optical setup or a many-body ultracold atom system in a certain limit.

Upon measuring e.g. the Husimi-Q distribution of the system, one can compute physical quantities that are of quantum-informational relevance, such as entanglement entropies. Yet, usually one is faced with the challenge that experimental measurements are limited by factors like shot size, noise etc.

In my project, we aim to develop Neural Quantum States, that can efficiently reconstruct quasi-probability distributions of continuous-variable, quantum many-body systems, ultimately to be able to do quantum “distributional” tomography from sparse and noisy measurement data.