Abstract
We demonstrate how machine learning is able to model experiments in quantumphysics. Quantum entanglement is a cornerstone for upcoming quantumtechnologies such as quantum computation and quantum cryptography. Ofparticular interest are complex quantum states with more than two particles anda large number of entangled quantum levels. Given such a multiparticlehigh-dimensional quantum state, it is usually impossible to reconstruct anexperimental setup that produces it. To search for interesting experiments, onethus has to randomly create millions of setups on a computer and calculate therespective output states. In this work, we show that machine learning modelscan provide significant improvement over random search. We demonstrate that along short-term memory (LSTM) neural network can successfully learn to modelquantum experiments by correctly predicting output state characteristics forgiven setups without the necessity of computing the states themselves. Thisapproach not only allows for faster search but is also an essential steptowards automated design of multiparticle high-dimensional quantum experimentsusing generative machine learning models.