Learning the interpretable representation of quantum entanglement, the depth generation model can be directly applied to other physical systems

Quantum physics experiments have produced interesting phenomena , Such as interference or entanglement , This is the core feature of many future quantum technologies . The complex relationship between the setup structure of quantum experiment and its entanglement properties is very important for the basic research of quantum optics , But it's hard to understand intuitively .

The University of Toronto cooperates with researchers from the University of science and technology of China , A depth generation model for quantum optical experiments is proposed , The variational automatic encoder is trained on the data set of quantum optics experiment . In a series of computational experiments , The team studied their quantum optical variational automatic encoder （QOVAE） Learning representation of and its internal understanding of the quantum optical world .

Experimental proof QOVAE We have successfully learned the interpretable representation of quantum optical experiments and the relationship between experimental structure and entanglement . The researchers showed QOVAE Novel experiments can be generated for highly entangled quantum states with specific distributions that match their training data .QOVAE You can learn to generate specific entangled states , And effectively search the experimental space for generating highly entangled quantum states .

what's more ,QOVAE How to construct its potential space is explicable , It can help find strange patterns that can be explained by quantum physics . Sum up ,QOVAE It can be applied to other physical systems immediately .

The research 「 Learning interpretable representations of entanglement in quantum optics experiments using deep generative models」 entitled , On 2022 year 6 month 16 Published on 《 Nature Machine Intelligence》.

Thesis link ： https://www.nature.com/articles/s42256-022-00493-5