ConQuER Controlling Quantum Experiments with Reinforcement Learning

Summary

This proposal aims to advance the integration of state-of-the-art machine learning and artificial intelligence (AI) with quantum physics experiments. The abundance of physics data coming from experiments and simulations of quantum systems allows us to use the power and efficiency of machine learning methods to extract information from this data in a way that goes beyond traditional methods. Based on this insight, I will design and implement a reinforcement learning (RL) method that can directly control, stabilize (to create a quantum memory) and tune (for device characterization and setup) a multiple-spin-qubit experiment at the Niels Bohr Institute in Copenhagen. The resulting framework and open-source AI software are expected to be useful in any other quantum experiment for which tuning is a major component, and is expected to generate a large impact in the community. Automatic device tuning will free up precious time resources that can be invested in significant experimental advances.

Resources Show all and search (5)

Unfold all

Fold all

More information & hyperlinks

Web resources:	https://cordis.europa.eu/project/id/895439
Start date:	01-10-2020
End date:	30-09-2022
Total budget - Public funding:	219 312,00 Euro - 219 312,00 Euro